TruptiPandya

An AI recruitment assistant that ingests CVs and job descriptions, generates semantic embeddings, and performs retrieval-augmented candidate-to-role matching. Multi-stage prompt pipeline with context design and output guardrails for auditable, grounded answers.

“End-to-end RAG product demonstrating prompt engineering, vector retrieval, and responsible AI practices — aligned with enterprise LLM deployment.”

AI-driven career coaching platform using LLM orchestration to deliver personalised skill assessments, learning roadmaps, and job-fit recommendations. Multi-turn conversational flows with structured context management.

“Demonstrates ability to orchestrate LLM workflows, apply responsible AI design, and communicate AI capabilities through intuitive UX.”

Medical screening assistant chatbot that guides patients through structured triage questions and recommends appropriate hospital wards. RAG-powered retrieval surfaces hospital-specific clinical guidance, while nurses retain full oversight and authority to override any suggestion.

“Responsible AI in a high-stakes healthcare context — grounded responses, human-in-the-loop control, and domain-specific RAG in a safety-critical workflow where accuracy and auditability are non-negotiable.”

GenAI-powered digital forensics tool — investigators upload evidence files (logs, reports, metadata), ask natural-language questions, and receive RAG-grounded answers with automated timeline reconstruction, anomaly detection, and downloadable structured reports. Per-case isolated FAISS indexes prevent cross-case data leakage.

“End-to-end RAG system with explainable findings, per-case conversation memory, and structured report generation — advanced prompt engineering in a domain demanding accuracy, citation, and full auditability.”