Anshuman Pandey

AI Engineer

India 0+ yrs exp 82 · Excellent

About

Performance-driven AI Engineer specializing in production-grade GenAI and RAG pipelines. Expert in reducing inference latency by 35% and scaling architectures for 10K+ users. Specialist in LangChain, FastAPI, Vector DBs (FAISS/Pinecone), and Edge AI deployment.

Skills & Expertise (33)

LangChain Intermediate

8.1/10

Years Exp

Semantic Search Intermediate

8.0/10

Years Exp

FAISS Intermediate

7.9/10

Years Exp

Python Intermediate

7.8/10

Years Exp

FastAPI Intermediate

7.7/10

Years Exp

Typescript GraphQL REST RBAC Real-time Inference Multi-tenant Architecture Scalable System Design AWS GitHub Actions Kubernetes Docker Php Redis Postgresql Go LangGraph RAG Transformers CNNs TensorRT OpenCV YOLOv8 TensorFlow PyTorch RAGAS vLLM Prompt Engineering Pinecone

Work Experience

Associate AI Engineer

Weber Innovations

Oct 2025 - Present

Spearheaded architecture of 5+ production AI platforms using FastAPI, increasing velocity by 40% and scaling to support 10K+ monthly active users with 99.99% service availability. Engineered a scalable multi-tenant backend using PostgreSQL and Docker, optimizing B-Tree indexing to reduce API latency by 45% while maintaining SOC2 compliance for secure data handling. Orchestrated enterprise CI/CD pipelines via GitHub Actions, implementing Redis caching to reduce P95 latency by 35% across high-traffic endpoints and cutting deployment effort by 60%.

AI Development Intern

Mawai Infotech Ltd.

Jun 2025 - Oct 2025

Deployed LLM-powered semantic search using LangChain, achieving a 28% gain in retrieval accuracy for 500+ users via hybrid retrieval and multi-vector search strategies. Architected production RAG pipelines utilizing FAISS and HNSW indexing, lowering LLM hallucination rates by 40% and reducing manual workflow times for cross-functional teams by 25%. Awarded Pre-Placement Offer (PPO) for delivering AI features that resulted in a measurable 15% business efficiency increase through implementation of automated context-aware decision systems.