Pankaj

AI Engineer with hands-on experience designing, building, and deploying production-grade LLM systems and agentic workflows. Built 3 end-to-end AI systems: a multimodal financial RAG platform with hybrid retrieval and CI quality gates (0.91 faithfulness / 89% Recall@5); a fine-tuned 7B model achieving 84.1% SQL execution accuracy against complex enterprise schemas; and a 5-node LangGraph multi-agent system with 91% error recovery. Strong backend foundation across FastAPI, PostgreSQL, Redis, and Celery with 2 years of production system ownership; deep practical fluency in LangChain, LangGraph, RAG pipelines, vLLM, and LLM observability.

Software Developer (Intern → Full-time) at Ireo Private Limited (2024-02 – Present)

• Architected a Visitor Management System (Django, PostgreSQL, Next.js) with QR-based check-in, role-based access control, approval workflows, and audit logging.
• Architected an HRMS platform (FastAPI, PostgreSQL, Next.js) serving 500+ daily active users with sub-200ms p95 API latency, JWT authentication, and RBAC across 12 permission levels.
• Designed audit-trail architecture spanning 20+ database tables and 3 PostgreSQL schemas, ensuring 100% data traceability and zero data integrity incidents.
• Built Celery + Redis payroll, attendance, and reimbursement automation, saving 10+ HR hours/week while maintaining a job failure rate below 0.5%.
• Owned HRMS production operations, reducing MTTR from 2 hours to 20 minutes via monitoring, alerting, and incident runbooks.

B.Tech in CSE – Hindu College of Engineering, Sonipat (2020-07 – 2024-06)

Data Science Mentorship Program – CampusX