Yash Munshi

Evaluated 20+ Large Language Model (LLM) responses within RLHF pipelines, improving response quality, factual consistency, and safety alignment across evaluation workflows. Performed 50+ structured analysis of model responses to identify edge cases, hallucinations, logical inconsistencies, and failure patterns. Collaborated on prompt evaluation and AI quality assurance workflows.

LLM Evaluation Engineer - Freelance (Project Groot) - Soul AI - Delaware, United States (Remote)

(2026-04)

• Evaluated 20+ Large Language Model (LLM) responses within RLHF pipelines, improving response quality, factual consistency, and safety alignment across evaluation workflows.
• Performed 50+ structured analysis of model responses to identify edge cases, hallucinations, logical inconsistencies, and failure patterns.
• Collaborated on prompt evaluation and AI quality assurance workflows.

Backend Developer Intern - NexCell Solutions - London, United Kingdom (Remote)

(2026-03)

• Developed 15+ scalable REST API endpoints using FastAPI and PostgreSQL for ConneX AI, enabling real-time workflow automation and backend integration.
• Built an event-driven automation engine using Celery and Redis for asynchronous trigger-action execution, improving workflow scalability and enabling reliable background task processing.
• Implemented DAG-based workflow orchestration with conditional execution logic to support extensible multi-node automation pipelines.
• Configured CI/CD pipelines using GitHub Actions to automate testing and deployment processes.

B. Tech - Electronics and Computer Science Engineering - Kalinga Institute of Industrial Technology (2023-07 - 2027-07)