AI + Full-Stack Engineer · Ahmedabad
I build grounded AI systems and full-stack products — from research experiments to applications with real users.
Selected work
Research, engineering, and the work that brings them together.
A self-improving scientific research agent. It fetches and validates ArXiv papers, stores paper atoms in PostgreSQL with pgvector, and supports a multi-agent hypothesis loop.
Contribution Paper ingestion, structured storage, and a multi-agent hypothesis workflow.
Trained a 50M-parameter decoder-only transformer on synthetic reasoning tasks and built an evaluation suite that exposed memorization despite near-perfect in-distribution results.
Contribution Transformer training and out-of-distribution evaluation.
A local RAG assistant for 19 reinforcement-learning papers with PDF ingestion, FAISS search, IDF-weighted reranking, TinyLlama generation, FastAPI endpoints, and Docker support.
Contribution PDF ingestion, retrieval, reranking, and API delivery.
An interactive study assistant with textbook-grounded chat, voice input, test generation, chat history, and a separate RAG engine for AI answers.
Contribution Textbook-grounded chat, voice input, and test generation.
A learning-first GPT implementation with attention heads, multi-head attention, transformer blocks, tokenizers, KV cache, grouped-query attention, training, and generation.
Contribution A from-scratch GPT implementation and training workflow.
A Streamlit NLP app that classifies news text as real or fake using TF-IDF features, Logistic Regression, confidence scores, and a labeled Kaggle dataset.
Contribution An NLP classification app with confidence scores.
Experience
A practical engineering profile: ship the product, learn the theory, then explain it clearly.
Work experience
AST Consulting
I build and maintain features for a production CMS/news platform, working across backend APIs, database models, and frontend functionality. I develop image-processing features, profile and cover image APIs, multilingual content workflows, and category and user-management functionality. I work with Node.js, MongoDB, Redis, Elasticsearch, and Docker, while debugging issues and improving API reliability. I also test APIs and features extensively using Postman, including role-based permissions, edge cases, and regression scenarios.
AUM Imagineering
Education & leadership
GEC Bhavnagar
CGPA 8.51/10 with focus areas in DSA, DBMS, linear algebra, probability, machine learning, and deep learning systems.
Stack
Research voice
Questions from my experiments with retrieval, reasoning, and evaluation.
The small-transformer study reports near-perfect in-distribution accuracy but complete exact-match failure on 7,137 out-of-distribution examples.
The Scientific Copilot direction turns paper overload into a structured system for ingestion, clustering, hypothesis generation, critique, and iteration.
AI Tutor and RetrievalStack connect retrieval pipelines to student questions, RL papers, voice interaction, test generation, and cited answers.
Available for AI, backend, and full-stack roles