Anushka Punekar

I’m an AI Engineer trainee focused on building and deploying production-grade LLM systems—especially RAG pipelines and agentic AI workflows. I work directly with clients and bring ownership across the full lifecycle, from requirement gathering to production delivery.

In my Agentic AI Developer Trainee role at ShiftAhead, I independently architect, build, and deploy end-to-end agentic AI systems across multiple concurrent projects. I gather requirements, propose AI-driven solutions, iterate on agent workflows, and manage delivery end-to-end without a dedicated engineering team.

I’ve implemented evaluation pipelines to assess relevance, correctness, and reliability, and I’ve strengthened observability with structured metric tracking, event-driven alerting, and self-healing agentic workflows. I also built backend pipelines for telecom KPI metrics (ASR, ACD, failure rates) using Python, FastAPI, and LangGraph.

Earlier, I completed a Cybersecurity internship at Goldman Sachs through the Tata Global Micro Internships Program, where I performed password vulnerability analysis with Hashcat and recommended stronger hashing standards (PBKDF2/bcrypt). My projects reflect the same mindset: dataset-driven LLM evaluation, hallucination detection, and multi-agent RAG systems for measurable retrieval improvements.