Sign in to view Balazs’ full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Sign in to view Balazs’ full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
London Area, United Kingdom
Sign in to view Balazs’ full profile
Balazs can introduce you to 10+ people at Veolia
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
368 followers
378 connections
Sign in to view Balazs’ full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Balazs
Balazs can introduce you to 10+ people at Veolia
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Balazs
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Sign in to view Balazs’ full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
About
Concerned about the…
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
Activity
368 followers
-
Balazs Dukat posted thisA few months ago while working in an early stage startup (or let's call it a partnership, to use the correct legal definition) I had a couple of ideas: 1. What if end-users could train or at least fine-tune their own local LLMs from plain text? Ideally starting from a very minimalist bloat and bias free model, which then could learn from text provided to it by the user, in a one-shot kind of way, without any resource-intensive training proces. 2. Even better, what if such an LLM could hold onto information and rules given to it by the user through regular interaction, holding it in context (assume a sufficiently large context window of many thousand tokens), and then it could be sent into a "sleep" (actually, learning and pruning) state where it could learn, and possibly "unlearn" (forget) contradicting rules and info. These could democratise LLM training and use, organisations and even individuals could train their own specialist local models. Apart from democratising customisation, another potential advantage: I think a limited and specialised model might be less likely to hallucinate. It could be more reliable. Less may be more. And I left the best part last: Alignment. Starting from something small and simple, and then training it the right way could yield a "good" LLM, rather than a soggoth with a mask on. A minimalist model trained by the right people could be free of unwanted knowledge, biases, and attitudes. I wonder if Hugging Face could support this. #LLM #AI #AIdemocratization
-
Balazs Dukat shared thisMy thoughts on this, and mine alone: 1. The internet in its present form is clearly not a safe space for children. So no internet acces for little ones. That is our family policy. 2. The manosphere should be considered in the same category as extremist groups. Because it is. Gender-based extremism. That should be government policy. 3. If your child knows of genders, has words for "girl" and "boy", then that child is old enough to be told about similarities, differences, and also to be taught empathy, maybe through examples and real-world interaction. I seriously wonder in what weird gender-segregated world can boys so overmystify girls that fall for the manosphere lies, have they never talked to a girl before?Balazs Dukat shared thisHOW MANOSPHERE IS LIKELY TO AFFECT THE FUTURE OF MANY CHILDREN While presented as a "Man Space"; its fastest-growing audience today is children and teenagers of all gender. The manosphere is a network of online communities centered on masculinity, gender roles, and the belief that society is biased against men. It includes groups like Men’s Rights Activists (MRAs), Pick-Up Artists (PUAs), Incels, and Men Going Their Own Way (MGTOW). While these communities differ, many promote rigid, adversarial views of gender. And children are increasingly being exposed to these ideas: 📌 Why This Matters for Children This is not just about online content. It’s about how children learn to see themselves and each other. 1. Identity Formation Is Being Influenced Early Children are encountering ideas like: “Alpha vs Beta” “Red pill vs Blue pill” 👉Worth tied to dominance, looks, or control At a developmental stage where identity is still forming, this can: ❎Distort self-worth (for boys and girls) ❎Reinforce harmful gender stereotypes ❎Limit emotional expression and empathy 2. Algorithmic Exposure Is Quiet—but Powerful Platforms like TikTok and YouTube often recommend content based on engagement. A child searching for: 👉 fitness 👉 confidence 👉 self-improvement …can quickly be led into increasingly extreme manosphere content. This “pipeline effect” is not accidental 📌it’s algorithmic. 3. It Shapes How Boys Relate to Girls—and Vice Versa Some manosphere narratives frame relationships as: Competition Control Manipulation This can lead to: ❎Increased hostility or mistrust between genders ❎Normalization of disrespect or emotional detachment ❎Reduced capacity for healthy relationships later in life ❎📍Girls internalizing these dynamics or adjusting behavior to navigate them. (Real-World Signals We Cannot Ignore) 👇 -The Isla Vista killings were linked to incel ideology and misogynistic beliefs. -The Toronto van attack was carried out by an individual who explicitly referenced incel communities. These are extreme cases—but they highlight a broader issue: 📌Online narratives can and do shape offline behavior. The Role of Influencers Figures like Andrew Tate have amplified manosphere ideas to millions of young followers, often mixing: Self-improvement advice Wealth narratives Anti-feminist messaging For a child, this can be confusing: What looks like “motivation” can carry deeply harmful beliefs underneath. So What Does This Mean for the Children's Future? If left unaddressed, we risk raising a generation where: •Boys feel pressure to perform rigid, dominant identities •Girls navigate increasing digital misogyny This makes it a childhood development issue. So I ask again: Who is shaping the digital environments children grow up in? What values are being embedded through algorithms? #ChildRights #DigitalSafety #OnlineSafety #Algorithms
-
Balazs Dukat shared thisThis looks interesting in a very good way, and lines up with something I had in mind. #carbon #climatechangeBalazs Dukat shared thisNetwork, Why Verra V5.0 is a tailwind, not a risk Verra’s VCS 5.0 is a reset for the voluntary carbon market. Higher bars on additionality, baselines, and auditability will reduce volume and improve quality. That is constructive for capital. The biggest impact is on avoided emissions. Avoidance can scale quickly, but only where the counterfactual is provable and conservative. V5.0 makes that non-negotiable. This is where DCAS Tech sits. DCAS is built for measured carbon avoidance: Baselines derived from live operational data Interventions linked directly to asset behaviour Conservative attribution of avoided emissions Full audit trails aligned to verifier and registry scrutiny Because the data is continuous, attributable, and non-duplicative, outputs can support high-integrity offsetting claims and not narrative-led ones. Commercially, this matters. We are currently engaged across a live pipeline of energy and infrastructure assets, progressing pilots and structured discussions with asset owners where: Digital twins or predictive maintenance systems are already deployed or deploying. Avoided emissions can be evidenced at asset level. Carbon integrity is a prerequisite for investment or reporting. Removals remain essential for long-term net zero. But in the near to medium term, credible avoidance is how scale is achieved. V5.0 doesn’t constrain the market. It reallocates value toward platforms that can evidence impact. You cannot get away with those quarterly spreadsheets, built on assumptions. V5.0 is about verification. We are building this business in the UK. And we are confident that it will change the face of ESG Reporting, and create a new class for Carbon Trading. Contact me for discussions about how we can help. B2B.
-
Balazs Dukat shared thisPhoto and transcript of some of my old notes. Persons' and companies' names have been redacted. 16/10/2019 Carbon offset scheme Operation Clear Sky Get |major client| to use our CO2 offset? send email to: |Names redacted|+|Names redacted| |Redacted| reply was somewhat dismissive, I keep pushing on and asking what we have, |client's website| is the top candidate, not |major client| itself. |Names redacted| weighed in to diffuse and support. - What can CO offset do? - What do we offer to clients? - What was done? We have CO2 calculator, and CO2 offset, by |LEGEND| Those need to be advertised by Accounts and Sales. (Supposedly in progress since 01/11/2019.) [What would happen if CO2 offset become compulsory?] 13/11/2019: There is a "product" being reviewed 20/11/2019: I requested to become the manager of this. |Redacted|'s reply: "Let's put it on hold..." "we could talk about a lightweight approach that you could take"
-
Balazs Dukat posted thisI invite my network (and those who are yet to connect) to join me in my initiative. I intend to start a group of professionals to devise and complete projects. This would serve to practice existing skills, learn new ones, and gain experience. From my part this would be centered on data projects, ideally with a focus on the environment or some other honourable cause, such as DEI. Other skill areas and interests are welcome. I can't offer pay, but I hope this would help us find paid employment, or possibly this could grow into something. DM me if you are interested, and I will come up with the next steps, for example, we might need a Slack workspace and a Github repo.
-
Balazs Dukat shared thisMy course completion certificate for: "dbt (Data Build Tool) : dbt for Data Engineers"
-
Balazs Dukat shared thisNature security assessment on global biodiversity loss, ecosystem collapse and national securityNature security assessment on global biodiversity loss, ecosystem collapse and national security
-
Balazs Dukat reposted thisBalazs Dukat reposted thisLet me tell you a story. About a little Black girl. Dark-skinned. Scrawny. Quiet. She didn’t have big visions of herself. Not because she lacked imagination —but because life kept interrupting it. Teachers who didn’t believe in her. Adults who didn’t protect her. Abuse that taught her to shrink. Messages that told her she was too much — and never enough. Dreams don’t always disappear loudly. Sometimes they’re just slowly pressed flat. But there was one thing that never left her. She wanted to help people. When you’ve been bullied. When you’ve been hurt. When you’ve felt abandoned, rejected, unseen — You develop a deep sensitivity to suffering. You notice pain faster. You recognise it in rooms before it speaks. And sometimes, you make a quiet vow: No one should have to feel this alone. I made that vow early. I decided — even as a child —that I would not pass on the pain I inherited. That I would do what I could to reduce harm, not replicate it. • That decision is why I became a social worker. • Why I trained as a psychotherapist. • Why I care so deeply about how systems affect bodies. • And it’s also why I care about money. • And AI. • And building businesses that work. • Not because I worship tech. • But because I want us to win. I love Black women. I was raised by a strong, brilliant Black mother. And while I know not all of us share the same experiences with Black women — I do know this: We deserve to flourish. We deserve ease. We deserve safety. We deserve access. We deserve futures that aren’t defined by survival. The system may not be designed with us in mind —but that does not mean scarcity is our destiny. When I teach AI, when I talk about building wealth, when I help Black women turn their expertise into income: I’m not chasing success. I’m honouring a promise. To that little girl. And to every Black woman who was told — explicitly or subtly — to dream smaller. We deserve more than survival. We deserve to thrive. ♻️ Repost this if you believe Black women deserve not just to survive systems but to outgrow them.
-
Balazs Dukat shared thisOne more reason for me to explore better ways of feeding data to models, I guess.Balazs Dukat shared thisWe Need To Talk About The Elephant In The Data Centre 😁 You notice sometimes your AI models going slow or just stop working? Everyone’s banging on about AI being the future. Revolutionary. Game-changing. “It’ll solve everything!” It's absolutely an amazing technology! Cool. Brilliant. But has anyone actually done the maths? Here’s the bit nobody wants to mention at dinner parties: training a single large AI model now uses roughly the same electricity as 5 average UK households consume in a year. A single model. And we’ve already got thousands of them. Meanwhile, we’re all clutching our reusable coffee cups and switching to LED lightbulbs like it’s going to offset checks notes the electricity grid struggling to keep the world’s AI running. The inconvenient truth: We’ve become obsessed with building smarter AI whilst barely discussing how to power it smarter. It’s like buying a Porsche and worrying about the fuel efficiency of the engine block. Here’s what actually needs to happen: We need to talk about energy as part of the AI conversation, not as footnotes in white papers → Efficiency matters more than raw capability (yes, sometimes less is more) → Renewable energy infrastructure for data centres isn’t a nice-to-have, it’s table stakes The real innovation isn’t just in the algorithms anymore, it’s in building AI systems that don’t require us to build a small nuclear reactor for each one. So let’s flip the script: instead of 'what can AI do?' ask “what can AI do sustainably? That’s the actually interesting question. What’s your take? Am I off my head, or does this deserve more airtime? ta-ta 🙃
-
Balazs Dukat reacted on thisBalazs Dukat reacted on this📉 Amikor a "nincs munka" nem pihenés, hanem fegyelmi csapda. (Az üresjárat protokoll csapdája) Ismerős az idegőrlő csend az irodában? Bemész, megcsinálod a feladatodat 2 óra alatt, és utána... semmi. Nincs új projekt, nincs e-mail, a főnököd elérhetetlen, te pedig ott ülsz az asztalnál, és telnek az órák. ⚠️ A munkavállaló végzetes hibája: Mivel unatkozol, az agyad automatikusan a dopamin után nyúl: Előveszed a telefont és görgeted a Facebook-ot/TikTok-ot, és abban a pillanatban belép a főnök. A tekintete azonnal megváltozik. Nem azt látja, hogy "a szorgalmas kolléga vár a feladatra". Hanem azt: "Maga mit csinál? Munkaidőben szórakozik? Ez a tisztelet hiánya és fegyelmi vétség." És a legfájdalmasabb igazság: Önként sétáltál bele a pofonba, mert igaza lehet (de nem mindig). Hiába ő a hibás, amiért nincs munka, te adtál neki fegyvert a kezébe azzal, hogy látványosan nem állsz rendelkezésre. 📉 A menedzsment erőforrás-pazarlása: Vezetőként ha a beosztottad a plafont nézi vagy telefonozik, az elsősorban nem az ő lustasága, hanem a te szervezési inkompetenciád. Fizetsz egy drága erőforrást (bér + járulékok), de nem hasznosítod, ez a Lean menedzsmentben a veszteség iskolapéldája. Ha nem monitorozod a terhelést, a dolgozó a pszichológiai "energiaminimum elve" alapján a legkisebb ellenállás felé (a telefonjához) fog nyúlni. Ez törvényszerű, így a te feladatod, hogy a potenciált munkává alakítsd, ne pedig hagyd elfolyni. 🧠 Zipf-törvénye / Energiaminimum elve: George Kingsley Zipf nyelvész fogalmazta meg: Minden emberi viselkedés a "legkisebb erőfeszítés elve" felé tart. Ha a rendszer (a munkahely) nem ad strukturált ingert (feladatot), az agy keresni fog egy könnyen elérhető, alacsony energiájú dopaminforrást (görgetés). Ez egyszerű biológia, vezetőként pedig ezt nem "büntetni" kell utólag, hanem megelőzni feladatkiosztással. ⚖️ A jogi pajzs (Mt. 146. §): Ha a munkáltató nem ad munkát, az ÁLLÁSIDŐNEK minősül. A törvény logikája: Te ott vagy, eladtad az idődet, ezért neked alapbér jár. (Igen, azért fizetnek, hogy ülj és várj!). DE! Az állásidő alatt is "rendelkezésre kell állnod". Ez a kulcsszó. A TikTok-ozás jogilag nem rendelkezésre állás, hanem a munkaidő jogellenes felhasználása, ha úgy ítélik meg. 🛡️ Az elhárítás: Hogyan védekezz munkavállalóként, ha az üresjárat csapdájába esel? Alkalmazd a "szakmai készenlét" stratégiát. - A monitor-szabály: Ha nincs munka, nyiss meg egy szakmai dokumentumot (PDF, Excel, belső szabályzat). Legyen a képernyőn "munka jellegű" tartalom. - Az önfejlesztés: Olvass szakmai cikket a gépen. Ha a főnök megkérdezi, mit csinálsz, a válasz: "Mivel várok az anyagra, addig a piaci trendeket elemzem, hogy naprakész legyek." - A zseb-protokoll: A privát mobilt hagyd a táskában. Ne adj nekik vizuális bizonyítékot a lógásra. Stratégiai tanulság: Ne engedd, hogy a cég a saját szervezési hibáját (üresjárat) a te szankcionálhatóságoddá (telefonozás) keretezze át. Ameddig fizetnek, addig fejleszd a hatékonyságod a munkában 🛑
-
Balazs Dukat reacted on thisAs a woman who runs alone, often in the bush, I can tell you this is exhausting. Before I even step out the door, there’s a mental checklist: • Where am I going? • Will my phone have reception? • Should I avoid that track? • Who knows where I am? - What time does the sun come up? Even how I wear my hair becomes a safety strategy. If I’m alone, I plait my hair and tuck it into my cap so it can’t be grabbed. This is the reality for so many women. We adapt, we plan, we limit ourselves and the places we can go own. We carry the responsibility. And while practical safety advice has its place, we need to be honest about the bigger issue. The burden should not sit solely on women to prevent violence. We need to be teaching something far more fundamental: Men do not own women’s bodies. Men do not have the right to intimidate, harass, control, or harm women. Objectification, harassment, rape, and violence are choices made by perpetrators and NOT consequences of women’s behaviour. Yes, let’s talk about safety. But let’s also shift the conversation to education and accountability. Because I, and every other woman, deserve to run, walk, live, and exist in the world without having to strategise our safety at every step. Men, before you go for a run, what checks do you do to keep yourself safe? Do you wait for daylight so you’re not running in the dark? What goes through your mind when you see another man on a remote track? Do you avoid certain routes because there’s no phone reception? Do you keep your phone in your hand just in case? Or is the only real preparation for the trails making sure you’ve packed a snake bandage? I know there are many good men out there. But being “not violent” is not enough. Speak up. Call out disrespect, harassment, and minimising language when you hear it. Challenge other men, especially in spaces where women aren’t present. Teach boys what respect actually looks like. Refuse to laugh along to jokes or content that objectify women Violence against women is not just a women’s issue. It is a men’s issue to confront. Change can come faster when men start standing alongside us and saying enough is enough Safer to be a female surfer with the sharks than a female runner on land
-
Balazs Dukat liked thisBalazs Dukat liked thisThe difference isn’t whether AI writes code. It’s whether you continuously direct architecture, constraints, and decisions.
-
Balazs Dukat reacted on thisBalazs Dukat reacted on thisI have a confession to make. Yesterday, I got paid to give a speech about how women can reclaim our true selves when we disappear, and I felt myself disappear a little. In that moment. I was so honored to be on that stage, and even more astounded by who else joined me - a Nobel peace laureate, a solo arctic explorer, an astronaut and rights activist, and several other insanely high-powered women. And as I walked on that stage, and after, a voice in my head said: "what on earth am I doing up here? I didn't do anything extraordinary." I spent the better part of yesterday fixating: 💙 on what went wrong 💙 on the pieces of my story I didn't get to finish telling 💙 on that one time I used the word 'jerk' and was it inappropriate? 💙 on whether I should have been up there at all But last night something extraordinary happened. I told my girls about my speech. I explained the core concept - which is that in 4th grade there was a specific moment where I felt myself fracture. A moment where my true, 'wild' internal self went underground, and my serious, 'good girl' external self took over. I told them that this happens to almost every woman I know. And that I'm on a journey to reclaim my wild self. She's just been in an alternate universe, hiding, waiting to be called back. I told them I have this idea to spend time with this idea. "What if I could bring that wild world to life, visually?" I asked them. What if you could help me? What if we could render the wild world where girls hide while women try to contort, shrink, fit in? What if you could stay in your wild instead of ever fracturing? My eldest went off and began drawing immediately. No questions. Just this small piece of perfection she handed back. My takeway: I suppose even I need to do a better job listening to the message I've been preaching. And, my daughters are apparently the ones who will help me do it. (and yes, this book project is real. it's already under way.)
-
Balazs Dukat liked thisBalazs Dukat liked thisMost careers are designed for people with invisible support systems. Yesterday, that theory joined my client meeting. We were discussing very important grown-up things: conversion rates, funnels, revenue... My son was discussing my husband’s secret chocolate stash. He had quietly snuck into the office, climbed into my husband’s chair like a tiny CEO, located the hidden chocolate (which, honestly, was news to me too), and started singing while eating it. I muted myself. We were mid sentence about growth and numbers. Then I had to speak again. I unmuted. Right on cue: “MAAAAMMMIIII!!!” Not a subtle mama. A full-commitment, chocolate-fuelled maaammmiiii. My client laughed. I laughed. But first… I felt embarrassed. Which is interesting, because objectively nothing unprofessional happened. No deadline missed. No strategy harmed. No revenue lost. Just reality… briefly unmuted. But many career models are still built around the assumption that someone else is quietly handling life in the background. Someone else absorbing interruptions. Someone else solving the urgent chocolate situations. So when life shows up loudly in the meeting, we feel like we broke a rule. Even when we know we didn’t. Because in the end: I am maaammmiiii. And I also talk about revenue. And both are part of the same reality. ————- I write about redesigning work for real lives, not idealized ones. Follow if that resonates.
-
Balazs Dukat reacted on thisBalazs Dukat reacted on thisWe need women to give birth. We penalize them when they do. We want more women in leadership. We also need women to have children. Otherwise there is no future workforce. No economic growth. No functioning society. And yet pregnancy is still treated like a surprise at work. It isn’t. Pregnancy is one of the most predictable transitions in life. We know roughly how long it lasts. We know roughly what the physical impact is. We know capacity changes. We know recovery takes time. We know reintegration matters. None of this is new information. And still: No real preparation before. Little meaningful adaptation during. Weak reintegration after. Women work almost until birth. Come back quickly. And are expected to perform exactly as before. As if creating a human had no impact on the body, the brain, or a person’s life. Then we look at leadership pipelines and ask: Where are all the women? 1 in 4 women leave the workforce within a year of having a child. Not because ambition disappears. Because the system was never designed for this reality. We say we want more women at the top. But we build organizations around a life model where pregnancy ideally doesn’t happen. Pregnancy isn’t unpredictable. The lack of preparation is a choice ——— I write about leadership, systems, and redesigning work to reflect how people actually live. Follow if you’re interested in building companies that don’t quietly push talent out.
-
Balazs Dukat liked thisAs a hiring manager then Open to Work banner doesn't say desperate, it doesn't say avoid, it says that person is Open to Work. It's that simple isn't?! If on the other hand someone does discriminate because the potential employee has the Open to Work banner on - imagine what else triggers them. They're the ones worth avoiding in the first place.
Experience & Education
-
Independent Consultant
********* ********
-
******
********* ******** * *** ********* *******
-
************
****** **** *******
-
********** ** *******
******** ****** ********* *********** 2:1 equivalent
-
View Balazs’s full experience
See their title, tenure and more.
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
or
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Licenses & Certifications
Volunteer Experience
Languages
-
Hungarian
-
-
English
-
Recommendations received
-
LinkedIn User
“I’ve had the pleasure of working with Balazs in the past at TF and more recently at an AI startup, and across both experiences his ideas and commitment to his work truly stood out. He consistently brought thoughtful, innovative perspectives to the table and wasn’t afraid to challenge the status of projects in many constructive ways. Balazs combines creativity with a strong work ethic, and he approaches every project with genuine passion and focus. He is dependable, forward-thinking, and a great collaborator, making him a real asset to any team. I would highly recommend him to any organisation looking for someone who brings both vision and dedication to their work.”
7 people have recommended Balazs
Join now to viewView Balazs’ full profile
-
See who you know in common
-
Get introduced
-
Contact Balazs directly
Other similar profiles
Explore more posts
-
Connor Powell
Southwark Council • 989 followers
🚀 ASC Data Hub – Now Live in Slough 🚀 This week we launched the Adult Social Care (ASC) Data Hub in Slough Borough Council 🎉 The ASC Data Hub brings together a single, trusted Power BI data model to deliver one version of the truth across Adult Social Care. This means: ✅ Centralised reporting ✅ Self-service access to data and insights ✅ Consistent, reliable KPIs Senior leaders now have key performance measures at their fingertips, supporting faster, more informed decision-making. At the same time, operational leads can drill down into detailed analysis and client-level data, enabling real insight into demand, performance and outcomes. This is a major step forward in how we use data to support practice, improve oversight, and ultimately deliver better outcomes for residents. David Coleman-Groom FCMI Fadzai Tande
51
6 Comments -
Bukola A.
Nationwide Building Society • 2K followers
Metadata in Action: Turning Tribal Knowledge into Organisational Intelligence 💡 Have you ever noticed that in many organisations real knowledge about data lives in people’s heads? You know the one person who knows what a particular field in the report actually means, or the only analyst who can explain why a number doesn’t match across systems. This is referred to as ‘Tribal Knowledge’, it’s valuable but fragile because when those people move teams or leave the company, that knowledge goes with them. This is where metadata management steps in. Metadata transforms that invisible expertise into shared organisational intelligence. ✅ It captures context — what the data means, where it comes from, and how it’s used ✅ It documents lineage — showing how data flows across systems ✅ It builds trust — because people understand why data looks the way it does When you manage metadata effectively, you’re not just managing data You’re preserving institutional memory and reducing dependency on a few key people. 👉 How is your organisation currently capturing and sharing data knowledge? #Metadata #DataGovernance #DataQuality #DataManagement #DataCulture #Analytics
20
2 Comments -
Abhisek Sahu
Rabobank • 153K followers
Some great platforms to practice SQL for free. 1️⃣ SQL Murder Mystery – https://lnkd.in/emNnuhgE 2️⃣ SQL Island – https://lnkd.in/e8YGFjzB 3️⃣ SQL Police Department – sqlpd.com 4️⃣ Schemaverse – https://lnkd.in/ec8gE5uT 5️⃣ Sqlzoo - https://lnkd.in/eJxHtRPN 💡 Save this post to use these resources later during your preparation. Follow Abhisek Sahu for more !! #sql
107
18 Comments -
Nick Holt
The BI School • 22K followers
'We were Tableau developers, and good ones at that, but we couldn't write our own SQL queries, so we were kind of limited.' That's what the analytics team at Leicester City Council told me about their ‘before’. They were: • Waiting for the database team to write queries • Unable to modify reports quickly • Stuck when requirements changed • Limited in what they could deliver After 10 weeks of training: • Creating their own queries • Modifying reports independently • Responding to new requirements same-day • Building more complex visualizations Sometimes the bottleneck isn't the tool - it's the gap in capabilities. Has your team hit a similar wall? #analytics #datavisualization #businessintelligence #sql #learnsql
8
9 Comments -
Newton School
128K followers
𝐄𝐯𝐞𝐫𝐲𝐨𝐧𝐞 𝐰𝐚𝐧𝐭𝐬 𝐭𝐨 𝐛𝐞 𝐚 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐭𝐢𝐬𝐭. But most end up doing an Analyst’s job without realising it. Before you join the hype train, understand what truly separates the two – the skills, the tools, the pay, and the kind of problems you’ll solve. If you’re serious about building a Data career that fits you, our 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐜𝐞 & 𝐀𝐈 𝐏𝐫𝐨𝐠𝐫𝐚𝐦 helps you get there step-by-step, with mentor guidance and real projects. Learn more: https://lnkd.in/gz_qyK6K #NewtonSchool #DataScience #NewtonDataScience
18
1 Comment -
Dipankar Mazumdar
Cloudera • 17K followers
I have been following the amazing work that has been going on in the Apache DataFusion Comet project 🚀 Yesterday, the 0.11.0 version was released. Highlighting some of these new/improved stuff. Key is to watch the 'type' of PR landing in the project to tackle a diverse set of things in the data systems space. ✅ Native Parquet Modular encryption support: - meaning even when Apache Spark offloads query execution to Comet’s native (Rust) runtime, it can still decrypt encrypted Parquet files securely - uses Apache Arrow-rs & Datafusion under the hood ✅ Improved Spark support: - version 4.0.1 - enhances ANSI mode support, adding full compliance for arithmetic, integral divide, rounding & remainder operations ✅ Apache Iceberg: - 1.9.1 version support - Adds a new API that removes dependency on Parquet’s ColumnDescriptor, resolving method-not-found errors and enabling smoother Iceberg-Comet integration - Fixes connection leaks observed in production Iceberg workloads ✅ Better Memory Management: - brings huge improvements to memory management, making it easier to deploy and more resilient to out-of-memory conditions Release notes in comment. Highly recommend following the project. #dataengineering #softwareengineering
114
5 Comments -
Shreya Arya
Virgin Media O2 • 11K followers
Date functions make it easy to analyse periods and spot performance cycles. Here are 8 TIPS to help you get started! 💜 This week your challenge is to use a Date Calculation in your visualisations! 📅 Explore data around the history of government shutdowns and see what stories the dates tell! 📖✨ Prompt Data: https://bit.ly/4qo7pK4 -- 💜 B2VB Submission Tracker: https://bit.ly/496vP13 🏠 Explore all our B2VB tips here: https://bit.ly/3YMsrGR 🩵 Tag our team: Eric Balash, Shreya Arya, Victory Omovrah #b2vb #tableau #datafam #dataviz #tips #date #analysis
72
5 Comments -
Alok Ranjan
VideoVerse • 42K followers
90% of Data Analysts don't know the thought process behind the tables they query. They work in silos, limiting themselves to just SQL and dashboards. But do you actually know why we need a data warehouse? or how the "Analytics Engineer" role emerged? To succeed today, you need to understand the full stack—from AI evals to data products. I made a video explaining the entire data lifecycle in de7tail, right from generation to final consumption. Master this to stand out in interviews and solve problems like a pro. Check out the video in hindi here: https://lnkd.in/g7bNz8NJ
16
-
Coursera
3M followers
Thinking about a career in data science? The demand for data science professionals is soaring, projected to grow by 36% by 2033. And now, you can earn your degree with The University of Huddersfield on Coursera! The hands-on,100% online BSc Data Science is built for success—flexible, affordable, and with support through lifelong access to the university’s career services. Plus, with Performance-Based Admissions, you can start your journey without traditional academic qualifications. Ready to start your data science journey? Learn more about the BSc Data Science from the University of Huddersfield: https://bit.ly/4joTYW5
83
4 Comments -
Yan Barros
BiaTech Corporation • 8K followers
DAISI: Data Assimilation with Inverse Sampling using Stochastic Interpolants Andrae, Larsson, Takao, Landelius, and Lindsten https://lnkd.in/d745_D-R This paper introduces DAISI, a novel data assimilation (DA) algorithm leveraging flow-based generative models to improve state estimation in complex, non-Gaussian systems. The core idea is to use a pre-trained, stationary generative prior (think normalizing flow) to represent the system's state distribution. Observations are then assimilated via a guidance-based conditional sampling scheme within the latent space of the flow. Crucially, DAISI incorporates forecast information through a clever "inverse sampling" step. Here's the technical breakdown: 1. Generative Prior: A normalizing flow, `p(z)`, is pre-trained on historical data to learn a representation of the system's state distribution. This prior remains fixed during the DA process. 2. Inverse Sampling: Given a forecast ensemble `x_f` (obtained from a numerical model), DAISI maps these forecast states into the latent space of the pre-trained flow using an inverse mapping `z_f = g(x_f)`, where `g` is the inverse of the flow's transformation. This provides initial conditions in the latent space for the subsequent conditional sampling. 3. Conditional Sampling: Using the latent forecast ensemble `z_f` as a starting point, DAISI performs conditional sampling to obtain an analysis ensemble `z_a` that is consistent with the observations `y`. This is achieved by minimizing a loss function that balances the likelihood of the latent states `z_a` under the prior `p(z)` and the likelihood of the observations `y` given the corresponding states `x_a = f(z_a)`, where `f` is the forward mapping of the flow. The paper uses a stochastic interpolant to guide the conditional sampling, ensuring a smooth transition between the forecast and the analysis. The key advantage of DAISI is that it avoids retraining or fine-tuning the generative prior at each assimilation step. This makes it computationally efficient and suitable for real-time applications. The inverse sampling step allows encoding model dynamics into the DA pipeline without requiring explicit knowledge of the model's equations within the flow architecture itself. --- Traditional DA methods often struggle with non-Gaussian distributions and high-dimensional systems, relying on approximations like the Ensemble Kalman Filter (EnKF) with heuristic tuning parameters. DAISI offers a more principled approach by learning a data-driven prior that can capture complex, non-Gaussian features of the system's state space. By decoupling the generative prior from the forecast model, DAISI provides a flexible framework for integrating machine learning with traditional numerical models. This is particularly relevant for applications where the underlying physics are complex and difficult to model accurately, such as weather forecasting, climate modeling, and subsurface flow.
26
-
Nicole Mark
Action • 3K followers
Do you have data "trapped" in Tableau .hyper files that you want to extract as JSON or to a .csv? Or do you have a .hyper file that you can't remember what's in it? (I had about 25 of those.) I made a little tool to inspect and extract the data using the Hyper API. You can use it on the command line or via web interface. It's simple, but it gets the job done! https://lnkd.in/eEDHbAtH
29
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More