“had the pleasure of working closely with Arijit during multiple AI product launches at quso.ai: including AI Influencer, AI Video Generator, AI Writer, and CutMagic. What impressed me most was Arijit's ability to translate complex infrastructure and backend systems into seamless user-facing functionality. Whether it was building scalable pipelines that could handle tens of thousands of video minutes daily or rapidly iterating on backend APIs based on user activation trends, Arijit consistently delivered with both precision and speed. His understanding of product impact, combined with technical depth, made him a critical partner in ensuring our go-to-market plans were not only ambitious but feasible. Beyond his technical capabilities, Arijit is deeply collaborative and user-centric. He routinely participated in product and growth discussions, contributing ideas that directly shaped what we built. If you're building 0→1 infrastructure or scaling a fast-growing AI product, Arijit is someone you'd be lucky to have on the team.”
About
Activity
-
You’ve got to love your story enough to tell it 150 times a week or sometimes more. To pitch it after a transatlantic flight. To smile through…
You’ve got to love your story enough to tell it 150 times a week or sometimes more. To pitch it after a transatlantic flight. To smile through…
Liked by Arijit Roy ▶️
-
Security has always been table stakes for us. Feather (YC S22) is now SOC 2 Type II compliant. This formalizes the security and reliability…
Security has always been table stakes for us. Feather (YC S22) is now SOC 2 Type II compliant. This formalizes the security and reliability…
Liked by Arijit Roy ▶️
-
🚀 Excited to announce that I’ve joined Net Square Solutions as a Security Analyst Trainee! Two weeks in, and I’m already learning a ton. From…
🚀 Excited to announce that I’ve joined Net Square Solutions as a Security Analyst Trainee! Two weeks in, and I’m already learning a ton. From…
Liked by Arijit Roy ▶️
Experience & Education
-
Feather (YC S22)
View Arijit’s full experience
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Licenses & Certifications
Volunteer Experience
-
President
Randomize MUJ
- 1 year 1 month
Science and Technology
Led the official computing club of Manipal University Jaipur.
- Speaker for technical events on topics like machine learning, competitive coding, cloud computing etc.
- Mentored over 350+ freshers in the field of machine learning and backend development
- Organized TechTales, the annual technical festival of Randomize();
- Lead organiser of Hacktoberfest(2020) Jaipur event -
Microsoft Learn Student Ambassador
Microsoft
- 2 years 5 months
Science and Technology
Learn Student Ambassadors are a global group of campus leaders who are eager to help fellow students, create robust tech communities and develop technical and career skills for the future.
https://studentambassadors.microsoft.com/studentambassadors/certificate/d9e56245-53b6-474f-9aa0-bb2745dd60c3 -
Speaker @ Hacktoberfest
Microsoft Learn Student Ambassadors
- 2 months
Science and Technology
Topics: What is open-source, how to start contributing to open-source
-
Open Source Contributor
Redis
- 1 month
Science and Technology
Open Source Contributor in Redis-developer Python project
-
Open Source Contributor
PyTorch
- 2 months
Science and Technology
Open Source Contributor in torch audio
-
Core Member
Amazon Alexa Community - Jaipur
- 1 year 2 months
Science and Technology
Publications
-
LyriX : Automated Lyrics Transcription Using Deep Learning-Based Residual UNet Model
IEEE
Automated Lyrics Transcription has emerged as a crucial research area to enhance music analysis and comprehension. In this paper, we present a novel approach for Automated Lyrics Transcription for English pop and rock songs using a deep learning-based residual UNet model. A comprehensive dataset of annotated songs, comprising lyrics and their corresponding audio tracks, was curated to train and evaluate the model. The performance of the vocal separation model was evaluated using the…
Automated Lyrics Transcription has emerged as a crucial research area to enhance music analysis and comprehension. In this paper, we present a novel approach for Automated Lyrics Transcription for English pop and rock songs using a deep learning-based residual UNet model. A comprehensive dataset of annotated songs, comprising lyrics and their corresponding audio tracks, was curated to train and evaluate the model. The performance of the vocal separation model was evaluated using the Signal-to-Distortion Ratio, demonstrating significant improvements over existing methods. Additionally, we propose an innovative lyrics transcriber, leveraging the vocal separation model, and evaluate its accuracy using the Word Error Rate. The results indicate high precision in the lyrics transcription process. Our research contributes to the advancement of automated lyrics transcription techniques, offering valuable insights for further developments in music analysis and application areas
Other authors -
Projects
-
rezonance
rezonance is a social music ecosystem targeted towards music enthusiasts. It consists of a music streaming application (available both for web & Android), a Discord bot, an eCommerce store and a multiplayer game. Currently, we have over 700 MAU.
- Built a pipeline of stages to query, aggregate and filter artist, album and music metadata from various data sources and joining them to store in a relational database.
- Built a custom search engine to personalize content searches for each…rezonance is a social music ecosystem targeted towards music enthusiasts. It consists of a music streaming application (available both for web & Android), a Discord bot, an eCommerce store and a multiplayer game. Currently, we have over 700 MAU.
- Built a pipeline of stages to query, aggregate and filter artist, album and music metadata from various data sources and joining them to store in a relational database.
- Built a custom search engine to personalize content searches for each user based on activity and preferences.
- Designed an end to end recommendation engine which auto tunes itself based on users interaction with the recommended songs.
- Developed micro-services in Python using FastAPI, PostgreSQL, Elasticsearch, Redis, Celery and containerized the application using Docker.
- Wrote unit tests using PyTest and created Ci/CD pipeline using GitHub ActionsOther creators -
-
Diode
-
Diode is an PaaS (Platform as a service) tool that lets people deploy their web-applications in two clicks.
This project was a runner-up in the Linode x Hashnode hackathon.
Currently, the frameworks supported are: Django, Flask, FastAPI, NodeJS, ReactJS, and NextJS.
- Deploed applications based on their run times automatically over SSH
- Used Celery to asynchronously handle shell commands while sending updates to the frontend about each step
- Used a reverse-proxy…Diode is an PaaS (Platform as a service) tool that lets people deploy their web-applications in two clicks.
This project was a runner-up in the Linode x Hashnode hackathon.
Currently, the frameworks supported are: Django, Flask, FastAPI, NodeJS, ReactJS, and NextJS.
- Deploed applications based on their run times automatically over SSH
- Used Celery to asynchronously handle shell commands while sending updates to the frontend about each step
- Used a reverse-proxy configuration using Nginx which also acted as a load balancerOther creators -
-
Amethyst
-
See projectAmethyst is a low-code, easy-to-use, GPU-powered recommender engine generator based on PyTorch. It requires only three parameters to rank/predict the best items for users and vice-versa
• Implemented BiVAECF & IBPR using PyTorch.
• Used Mean Average Precision, recall for top-k items for evaluation
• Applied stratified split and user-eligibility check for automatic pre-processing -
AQUA
-
Project AQUA is a Digital Evaluation Platform which leverages the power of leading-edge frameworks paired with state of the art artificial intelligence algorithms to provide an seamless and organized experience both for educators and students alike.
Other creators -
-
reZonance
-
reZonance is a Content Based Recommendation Service which suggests songs similar to the one chosen by the user. It has approximately 160,000 songs in its database and takes about 1500ms to generate recommendations.
Other creators -
-
HealthBridge
-
The aim of Project HealthBridge is to make the interaction of patients and doctors seamless, convenient and more organized. This platform is built keeping in mind the problems and hassles one has to face while using any medical facility via hospitals.
🤒 Here's what we provide for patients:
📄 Doctor appointment booking
🔬 Lab test booking
🧾 Online prescription & medical history database
💊 Generic drugs to brand name converter & recommendation…The aim of Project HealthBridge is to make the interaction of patients and doctors seamless, convenient and more organized. This platform is built keeping in mind the problems and hassles one has to face while using any medical facility via hospitals.
🤒 Here's what we provide for patients:
📄 Doctor appointment booking
🔬 Lab test booking
🧾 Online prescription & medical history database
💊 Generic drugs to brand name converter & recommendation system
🚑 Ambulance on demand service
🧬 Disease prediction services
🧑⚕️ For doctors, we provide the following services:
💻 Appointment management system
✒️ Online prescriptiion creation system
⚕️ View medical history and previous appointments of patient
❗ Prioritized lab results for disease infected patientsOther creators -
-
Pneumonia Detector
-
See projectPneumonia Classifier uses Deep Learning to classify X-Ray chest images to classify whether a person is infected with Pneumonia. Further it also can classify the type of infection i.e. whether the infection is viral or bacterial.
-
Automatic License Plate Recognition System
-
See projectALPR for India detects license plates from images of vehicles (upto 10 License Plates in a single image) and converts license plate numbers to strings. This can be used by law enforcement vehicles to scan, in real-time, vehicle insurance expiry, pollution check certificate validation, and possible blacklisted vehicles on the road. On a GPU powered machine, each detection takes approximately 70-80 milliseconds
-
Facial Expression Recognition
-
See projectFacial expression recognition aims to detect the persons feeling by using a video stream from the webcam. This detector can detect 7 human emotions which are Angry, Disgust, Fear, Happy, Neutral, Sad, Surprise.
-
Stock Predictor
-
See projectThe project aims to predict the stock prices of a given company using LSTM networks
Honors & Awards
-
Runner Up in AWS x Hashnode Hackathon
Hashnode
Vinyl is a multiplayer song guessing game inspired by the famous online multiplayer game, skribbl
-
Runner Up in Build with Linode Hackathon
Linode
Our project - Diode, a PaaS tool to deploy web-apps to Linode in 2 clicks won runner up in the Build with Linode Hackathon
-
Special Mention in Build The World with Dolby.io Hackathon
Dolby.io
Won special mention in Build The World with Dolby.io Hackathon, an international hackathon aimed towards building spectacular and innovative hybrid audio and video conference experiences
-
Placed 5th out of 300 teams in HackTheMountains2.0
HackTheMountains
HACK THE MOUNTAINS 2.O” was a 36 hour virtual Hackathon by the First ever Tech community from Jammu And Kashmir “SUDAN’S TECH ” - by the students and for the students.
-
Placed 1st in Coppell AI Hacks for Education
Coppell AI
Placed first in an educational themed international hackathon
-
Global Rank #1
CodeChef
Ranked 1 in May Long Challenge in Division 2
-
Microsoft Learn Student Ambassador
Microsoft
Learn Student Ambassadors are a global group of campus leaders who are eager to help fellow students, create robust tech communities and develop technical and career skills for the future.
https://studentambassadors.microsoft.com/certificate/d9e56245-53b6-474f-9aa0-bb2745dd60c3 -
Placed 3rd in HacksOverflow Hackathon
ACM-MUJ | IEEE SB MUJ
Placed 3rd in a National level Hackathon
-
TMA Pai Engineering Scholarship
Manipal University Jaipur
-
2nd Rank in Discovery Challenge
InterviewBit
-
TMA Pai Engineering Scholarship
Manipal University Jaipur
-
Silver Medal in Odyssey of the Mind (National Round)
Odyssey of the Mind
-
Gold Medal in Asia Pacific Conference of Young Scientists
FASE
-
Silver Medal in Odyssey of the Mind (National Round)
Odyssey of the Mind
Languages
-
German
Limited working proficiency
-
English
Full professional proficiency
-
Hindi
Native or bilingual proficiency
Organizations
-
Randomize MUJ
Head of Projects
- Present -
Metaverse (Cyber Society)
President
-
Recommendations received
2 people have recommended Arijit
Join now to viewMore activity by Arijit
-
Y Combinator hid a cookie under my chair 🍪 Last week at YC's sales conference, Claire Wu and I learned from John Banner that "anyone can…
Y Combinator hid a cookie under my chair 🍪 Last week at YC's sales conference, Claire Wu and I learned from John Banner that "anyone can…
Liked by Arijit Roy ▶️
-
🛑 Stop turning Audio into Text. It’s killing your latency. The standard Voice AI stack is broken: Whisper (ASR) → Text LLM → TTS. By the time your…
🛑 Stop turning Audio into Text. It’s killing your latency. The standard Voice AI stack is broken: Whisper (ASR) → Text LLM → TTS. By the time your…
Liked by Arijit Roy ▶️
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content