3+ years building production-grade AI systems — from RAG chatbots and NLP pipelines to Agentic AI — across Healthcare, Automotive, and IP domains. Pune, India.
I'm a Data Scientist and AI/ML Engineer with 3+ years of experience specializing in LLMs, Retrieval-Augmented Generation (RAG), and Agentic AI systems.
At CognoPI Solutions, I build end-to-end AI solutions for clients across Healthcare, Automotive, and Intellectual Property domains — taking projects from raw data ingestion to deployed, production-grade REST APIs.
I'm passionate about turning complex AI research into measurable business impact — whether that's reducing a 2-week patent search to milliseconds, or enabling 50+ corporate clients to self-serve health queries 24/7.
Built a medical RAG chatbot enabling clinic staff to query 100+ treatment guidelines (WHO, NIH, CDC, AIIMS) using PubMedBERT embeddings + Pinecone vector DB. GPT-4o-mini grounded responses to prevent hallucinations. Full CI/CD via GitHub Actions → Docker → AWS ECR → EC2.
End-to-end vehicle service intelligence platform ingesting 4 data sources (141K+ events) into a 120-feature dataset. Trained XGBoost models (F1=0.82, ROC-AUC=0.89) with SHAP explainability. Multilingual RAG chatbot in English, Hindi & Marathi.
Automated patent prior art search and CPC multilabel classification across 1M patent records (USPTO + EPO). Built a 7-step NLP preprocessing pipeline, PatentSBERTa embeddings, 4 similarity metrics, and an interactive Pyvis force-directed network graph.
Open to new opportunities, freelance projects, and interesting AI/ML collaborations. Feel free to reach out!