AI/ML Researcher & Production Engineer. I build reliable machine learning systems and study how to make AI safer, more capable, and useful in the real world.
Currently
M.S. by Research
IIT Madras
Data Science & AI
AI Safety
Secure AI Futures Lab
Technical operations & research
Developer
ML Hub
AI features & fixes
Top-Rated
Upwork
AI/ML Developer
I research privacy, harmful-response behavior, and robust AI systems, especially where multilingual and low-resource contexts matter.
From raw signals to monitored inference, I build systems that survive operational complexity, not just notebooks.
I design retrieval and multi-agent workflows with evaluation, guardrails, and an eye toward measurable reliability.
Much of my work is nodes and edges: knowledge graphs, retrieval, and multi-agent systems that reason over evidence. This one is alive.
Move your cursor, and click anywhere in it to grow the graph.
Open any case study for the technical approach, scope, and outcomes, not just a headline.
A research-grade system in development for multimodal deepfake detection in Indic-language contexts, combining specialized AI agents, evidence retrieval, and evaluation-driven safeguards.
Research in progress · India AI Mission
An evaluation harness that tests whether machine "erasure" is real across four layers of an ML stack: the database, the model weights, the vector index, and the semantic cache.
Two independent unlearning engines · one shared rubric
A cyclic multi-agent research system. Give it a claim; four agents argue it out live, re-searching themselves until confidence clears 80% or they run out of cycles.
Live WebSocket streaming · nothing precomputed
An operational 7-day power forecasting platform for distributed renewable assets. Built around per-plant learning, weather-aware features, robust ensembles, and automated inference.
~450 plants · 1,700+ models · 98% deployment success
An empirical study of privacy versus performance in recommendation-system unlearning across Matrix Factorization and SASRec architectures.
Evaluated privacy leakage, recommendation drift & stability
An interactive graph exploration system for dense professional-network data, connected to an enrichment, merging, and provenance-aware ingestion pipeline.
10k+ nodes · 5M+ raw records · 100+ source strategies
A personalized AI mock-interview experience that generates role-specific scenarios, scripts, voice-over, visuals, and interview simulations.
End-to-end product · BlockVerse Institute
An interview-feedback sentiment analyzer that turns interview text or video into structured scores, evidence, and actionable candidate feedback.
Whisper transcription · structured feedback reports
A context-aware mental-health platform that paired retrieval-augmented generation with safety guardrails and LLM optimization.
RAG lead & Llama Guard integration
An intelligent web search agent that validates queries, caches results semantically, scrapes top results, and generates AI-powered summaries.
Semantic caching · LLM summarization · Playwright scraping
Using few-shot learning, chain of thought, and advanced prompting techniques to assess and improve LLM performance in logical, analytical, and mathematical tasks.
Jadavpur University ML Hackathon 2024
A timeline of roles where I have helped turn AI ideas into research, systems, and outcomes.
Jul 2025 - Dec 2025
Academic backgrounds that shaped my research direction and engineering discipline.
Indian Institute of Technology Madras
Researching multimodal deepfake detection for Indic languages under the India AI Mission, guided by Dr. Krishna Pilutla and Prof. Balaraman Ravindran.
Heritage Institute of Technology
Graduated with a 9.22/10 GPA while building foundations across AI/ML, data systems, deep learning, and DevOps.
Tools serve the problem. My core work sits across AI research, production modeling, and systems engineering.
Kaggle has been part of my hands-on ML practice, from foundational regression work to competitive tabular modeling and iterative evaluation.
Visit Kaggle profileMLX Stage 1: Regression
Ranked 3 / 44
Regression · feature engineering
Obesity Risk Prediction
Ranked 84 / 3,587
Multiclass classification
Titanic: ML from Disaster
Competition work
Classification
House Prices
Competition work
Advanced regression
Selected certifications across AI, ML, cloud, and software engineering.
Building Multi-Agent Systems
CrewAI
Data Sage Hackathon
IIT Gandhinagar
ML Hackathon
IIT Gandhinagar
Function Calling and Data Extraction with LLMs
DeepLearning.AI
Prompt Engineering with Llama 2 & 3
DeepLearning.AI
Machine Learning
Stanford University / DeepLearning.AI
Neural Networks and Deep Learning
DeepLearning.AI
Generative AI
Google Cloud
Large Language Models
Google Cloud
Responsible AI
Google Cloud
Machine Learning on AWS
Amazon Web Services
Feature Engineering
Kaggle
SQL for Data Science
UC Davis
Intermediate Machine Learning
Kaggle
100 Days of Code: Python Pro Bootcamp
Udemy
Mastering DSA using C and C++
Udemy
I am putting together writing on AI safety, agentic systems, and lessons from shipping ML. First posts are on the way.
Let's build something meaningful
I am interested in research collaborations, thoughtful ML products, and teams working on reliable, safe AI. Let's talk.