Anemia detection from conjunctiva imaging
OpenCV + YOLO computer vision pipeline that cut clinical screening time by 40%, deployed as a live inference API at ICMR.
AI / ML Engineer · Delhi, India
I take machine learning from research notebooks to production.
I'm an AI/ML Engineer and Data Scientist currently building computer vision, generative AI, and retrieval-augmented generation (RAG) systems for healthcare research at ICMR headquarters — anemia detection at 92% accuracy, multimodal depression models, and production LLM pipelines that actually ship.
92%
CV anemia detection accuracy
40%
fewer LLM hallucinations via hybrid RAG
1,000+
inference requests served daily
9.4
CGPA · Gold Medalist, B.Tech CS
Selected projects
RAG research-paper assistant
Hybrid FAISS + BM25 retrieval-augmented generation (RAG) over academic papers with sub-second latency; 40% fewer hallucinations on internal benchmarks.
GPT-2 from scratch
Full generative AI language model architecture reimplemented in PyTorch — attention, BPE tokenization, weight loading — then fine-tuned for classification and instruction following.
Enterprise tool-calling LLM agent
An AI agent with 14 tools across 6 domains, multi-turn memory, and Pydantic guardrails — runs fully offline on local Llama 3.1 via Ollama.
Sentiment Sense — multimodal sentiment analysis
Led a 4-engineer team building a real-time multimodal deep learning system for emotion detection across text, images, video, and live webcam — CNN + Transformer ensemble at ~90% accuracy with REST + WebSocket streaming.
Experience
2025 — now
AI / ML Engineer · GA Digital Web Word
On-site at ICMR HQ, New Delhi — CV, multimodal deep learning, and data pipelines for national health research.
2025
Backend Developer Intern · Rough Note
Cut API response times 25% with query profiling, indexing, and Redis caching.
2024
Data Science Intern · Unified Mentor
End-to-end predictive modeling and time-series forecasting.
About
I'm an AI/ML Engineer and Data Scientist with 1+ year of production experience deploying machine learning in a real healthcare-research environment. My work spans computer vision (OpenCV, YOLO), NLP and LLMs (GPT-2, fine-tuning, RAG with FAISS + BM25 + LangChain), generative AI agentic systems with tool calling, and multimodal audio/speech models like Wav2Vec2 — all taken from research to deployed, production-grade APIs.
On the MLOps and backend engineering side: Python, PyTorch, FastAPI, Django, PostgreSQL, MongoDB, Redis, Docker, AWS, and GitHub Actions CI/CD. B.Tech in Computer Science (Data Science & AI) from Integral University, Lucknow — 9.4 CGPA, Gold Medalist, plus 11× IBM certifications in applied AI, machine learning, and data science.