Khushi Pawar

🚀 Aspiring AI/ML Engineer | I build AI systems that work beyond the notebook.

AI/ML Engineer who has shipped multilingual voice AI (English · Hindi · Marathi), LLM-powered conversational backends, RAG pipelines, and automated student assessment, built during my time at Thynk Technology and used by real users.

What I bring to a product team:
→ End-to-end ownership: architecture → backend → validation → deployment
→ GenAI & LLMs: RAG, prompt engineering, vector embeddings, LangChain, LangGraph
→ Backend engineering: FastAPI orchestration across 4+ model workflows, with API docs and runbooks the team could build on
→ ML rigor: RiskGuard AI, a fraud-detection platform with class-imbalance benchmarking, cost-sensitive threshold tuning, and SHAP-based reason codes for every prediction
→ Measurable delivery: ~40% faster cashier processing, 1,000+ assessments automated monthly, 60% less manual effort in RPA workflows

I don’t just fine-tune models. I design the systems around them: the APIs, pipelines, and integrations that let AI solve a real business problem.

Currently: extending into full-stack AI by learning JavaScript, React, and Next.js, so I can take a model from notebook to API to a usable interface, and building a portfolio of end-to-end projects.

Open to: AI/ML Engineer and Full Stack AI Engineer roles at product-based companies where intelligent systems are core to the product.

Let’s connect to talk about GenAI, applied ML, or opportunities.

Skills: Python · FastAPI · LLMs · RAG · LangChain · LangGraph · PyTorch · NLP · SQL · Docker · Azure · React.js & Next.js (learning)