Sandi Samantaray

"Can technology augment order writing capacity at regulators?" Unveiling article on The Leap Blog, of underlying paper, by Natasha Aggarwal, Satyavrat Bondre, Amrutha Desikan, Bhavin Patel and Dipyaman Sanyal of Trustbridge. https://lnkd.in/d62Ya9Vv

35

2 Comments

MIT’s latest research reveals a subtle yet critical flaw in transformer-based LLMs: “position bias.” These models tend to overweight the start and end of documents while overlooking the middle—potentially undermining accuracy in fields like financial compliance, legal review, and risk analysis. What can be done? MIT proposes solutions including: 1. Smarter attention masking 2. Refined positional encodings 3. Balanced training data As we scale AI at Northern Trust, we’re paying close attention to these foundational issues. Building trustworthy, context-aware AI starts deep in the architecture. #AI #ResponsibleAI #LLMs #FinancialServices #MIT #EnterpriseAI #GenAI #NorthernTrust #AITrust

33

RBI’s Digital Lending Directions 2025: A Step Towards Ethical, Transparent and Scalable Fintech:- The landscape of digital lending is evolving fast. With the RBI's latest Digital Lending Directions 2025, the ecosystem is being reshaped to bring sharper clarity, stronger borrower protections, and uncompromising transparency in fund flow, data privacy, and platform accountability. Over the last few days, I have studied the shift from the 2022 guidelines and compiled a practitioner-friendly compliance framework that breaks down the following: 1. What is new 2. What is mandatory 3. What is operationally complex 4. How institutions can stay audit ready without breaking agility From KFS restructuring to granular consent handling, from LSP governance to DLA accountability, this is not just about regulatory checklists. It is about rebuilding digital lending with trust, clarity, and long term sustainability at the core. I have shared the complete breakdown and a downloadable comparison table in this post. Would be keen to learn how others in the ecosystem are approaching their compliance and tech alignment for this directive. #RBI #DigitalLending #Fintech #Compliance #RiskManagement #NBFC #RegTech #DataGovernance #LendingInnovation

32

2 Comments

A particularly thought-provoking panel at the India AI Impact Summit 2026 brought together Paula Ingabire , Terah Lyons , and John Palfrey , three perspectives that neatly captured where the AI conversation is maturing. Paula Ingabire offered a powerful development lens: AI value cannot be measured purely in monetary terms. The real question is what problems we are solving, education, healthcare, citizen services, productivity. Capability building, closed-loop ecosystems, and national self-reliance are becoming foundational priorities. Terah Lyons highlighted a critical shift: The hardest AI challenges are no longer technological, but human and institutional. Responsible scaling, use-case-level risk evaluation, and global operating models will define how AI moves from experimentation to infrastructure. John Palfrey added essential balance: AI is not a separate or magical domain. Its long-term legitimacy depends on stable regulatory regimes and human-centered design, with civil society playing an active role alongside markets and governments. Across the discussion, one theme stood out: AI strategy is increasingly about systems, governance, and societal outcomes, not just models. The conversation is evolving, and so are the stakes. #IndiaAISummit #ArtificialIntelligence #ResponsibleAI #AILeadership #Innovation

15

1 Comment

Temenos Community Forum 2025: A quiet course correction? Last week’s TCF event offered a useful lens into how one banking technology provider is recalibrating its strategic focus. Not so much about unveiling bold new visions, but more about signalling a clearer understanding of where Temenos adds value… and where it likely doesn’t. The most significant shift? A pragmatic repositioning: ·      While the ambition to pursue large, globally established banks with entrenched core systems appears to have moderated, Temenos is now taking a more grounded approach: focusing its core banking proposition on mid-sized institutions, emerging market players, and digital entrants, while still offering a segue into larger institutions through targeted, specialist solutions that address specific add-on needs. ·      Internally, a leadership refresh has triggered a cultural rethink. Externally, the company’s messaging is now better aligned to the language of business stakeholders, anchored in ARR (Annual Recurring Revenue), market-relative growth, and operational delivery. ·      A standout moment came from EQ Bank | Equitable Bank. Their clarity on product engineering, automated testing, and digital execution – grounded in the bank’s lending DNA – offered a mature, actionable model for digital banking transformation. But there are still open questions. Has Temenos’ past technical edge, particularly in AI and AI explainability, been diminished, especially given its dominance at TCF 2024, fronted by Professor Hani Hagras, who co-founded Logical Glue, the explainable AI platform company acquired by Temenos in 2019? And with modular banking now a near-ubiquitous claim, what will distinguish Temenos as point solutions gain traction in Tier 1 and 2 spaces? The company’s focus on core performance and efficiency, reflected in its benchmark work with Microsoft Azure, is promising. As I’ve said elsewhere, such benchmarks are helpful indicators, but real-world impact will depend on deployment specifics and broader integration contexts. All in all, the direction of travel is clearer and more realistic. Execution, of course, will be the real differentiator. Curious how others are interpreting this shift, especially as our latest research shows that 67% of senior leaders now prioritise targeted IT investment over large-scale transformation. #TCF2025 #CoreBanking #BankingTransformation #Temenos #AIinBanking #DigitalBanking #FintechStrategy CCS Insight, Clive Howard

12

📖 Read the final instalment of our series on AI from the Strategy Monitor 🏦 Integration of generative AI is exposing the limits of banks’ legacy systems, with UK lenders spending £3.3bn each year maintaining outdated cores, sapping investment in AI-ready infrastructure. ☁️ Cloud integration and modular core systems are now viewed as critical enablers for maximising AI’s return on investment. Untangling years of “spaghetti” infrastructure remains one of the biggest hurdles. 🧠 As banks push ahead with their strategies, what are the greatest risks that could derail their progress? Read more: https://lnkd.in/eX5pQijp Andrii Degeler reports With insights from Jasmeet (Jas) Narang, Ian Glasner and Harriet Rees

24

Anthropic built a C compiler using a team of parallel Claude agents on a shared repo. This isn’t “AI autocomplete.” It’s multi-agent decomposition, parallel implementation, and integration/debug loops — i.e. agentic software engineering with real orchestration and review. The bottleneck shifts from codegen to coordination, control planes, and governance. Also a great choice of problem: a compiler is a perfect testbed for this approach — you can continuously validate output against existing compilers and specs. And yes, it’s written in Rust — which feels less like a coincidence and more like excellent taste in language selection 😉. Worth reading: https://lnkd.in/gd2zC3Ua #AIEngineering #AgenticAI #Rust #SoftwareArchitecture #DevTools #PlatformEngineering

20

2 Comments

It's always a pleasure speaking with Satish Acharya — his depth of expertise in financial crime and consent orders really shows. In this conversation, we dig into what clients are grappling with right now: policy standardisation, scaling teams, and how agentic AI and process redesign can drive real change. Worth a watch if you’re thinking about how to move beyond patching consent orders to building sustainable operations. #agentic #agentic_ai #banking #innovation #gen_ai #generative #genpact #banking_transformation #transformation #contact_center #customer_support #data #customer_support #gigafactory #commercial_banking #banking_operations #segmentation #fraud #agentification Lisa Galione Sachin Pai Manish Nayar Asim Burman Kavitha Shankaran Prakash Chacko Venkatachalam Narayanan Karun Aggarwal Eileen Ford Hedges, CAMS Priyanka Gaur Gaurav Agrawal Samir Saurav Alwin Bathija Moonish Sood Sameer Dewan Rohan de Souza Genpact Genpact EMEA Genpact Deutschland Dr. Eric Günter Krause (Ph.D. HSG) Ravi Alagh Vivek Shrivastava Vivek Sharma Daniel Bornstein Anu Sachdeva Rashmi Javed Sunod Kumar

104

2 Comments

🚨The National Payments Corporation of India did not have a transaction status check limiter in its technology architecture, which resulted in incessant API call requests by banks bringing down the UPI system on April 12, a couple of sources told Moneycontrol. NPCI runs the country's most popular digital payments platform, UPI. Application Programming Interface is a technology solution that helps two different institutions communicate. On April 15, Moneycontrol reported that NPCI, in its root cause analysis, found banks making non-stop transaction success checks continuously hitting the system beyond what was permitted, causing the platform to face downtime for around five hours, the longest in more than three years. According to an NPCI rule, the banks could check a transaction's success only thrice, and this instruction was to be set up by the banks and not NPCI, said one of the sources. As per NPCI operating circulars, the "Check Transaction" API call needs to be used only once in the interval of 90 seconds. Since this was the bank's responsibility, NPCI did not set a rate limiter at its firewall to stop API calls beyond three requests, said the second source. Imp one by Anand J https://lnkd.in/gbBwJMFf

22

HSBC demonstrates world’s first-known quantum-enabled algorithmic trading with IBM https://buff.ly/09RflYh HSBC has successfully executed the world’s first-known quantum-enabled algorithmic trading demonstration in collaboration with IBM. This groundbreaking achievement showcases the potential of quantum computing to revolutionize financial markets by enhancing trading algorithms with unprecedented speed and efficiency. Such advancements could lead to more precise market predictions and optimized trading strategies. As industries increasingly explore quantum technologies, this milestone highlights the transformative impact on global finance. For more details, visit: https://buff.ly/09RflYh. #QuantumComputing #AlgorithmicTrading

9

2 Comments