Himanshu Arekar

AI Engineer with 3+ years of experience designing and shipping production LLM and retrieval systems — RAG pipelines, agentic workflows, and evaluation frameworks — alongside hands-on experience owning full customer delivery: discovery, technical scoping, architecture, deployment, monitoring, and retention. Built and operate a production B2B marketplace-data API platform serving 300K+ successful requests per day at 99.8% client-verified accuracy, sub-280ms average latency, and 100% client retention across three retained accounts. Strong in Python, FastAPI, LLM application architecture, retrieval systems, multi-agent workflows, and production ML evaluation.

Founder & Technology Lead - NodusData HQ - Ujjain, India

(2025-01)

B2B Marketplace-Data Platform

• Architected and built a distributed data-extraction system (FastAPI, async queues, proxy infrastructure, autoscaling) processing 300K+ successful requests/day at 99.8% client-verified data accuracy and sub-280ms average latency.
• Designed a dedicated get_pc raw JSON structure API for core data-extraction requests, giving customers direct access to structured product/pricing data without post-processing.
• Founded and independently operate the platform end-to-end for three retained B2B customers requiring real-time marketplace data; own the full lifecycle from discovery and commercial scoping through architecture, delivery, and retention.
• Reduced infrastructure cost by 38% through bandwidth optimization and more efficient resource usage, while maintaining 100% client retention and a customer-required 99% minimum request-success SLA.
• Led a 10,000-request customer production proof-of-concept across Management, Finance, Product, and Technical stakeholders, validating accuracy, reliability under load, and regional compatibility before production approval.
• Authored OpenAPI specifications, SDK and integration guides, onboarding documentation, and SLA definitions.

AI/ML Engineer & Solution Architect - Strategic Execution Consultants Pvt. Ltd. - Indore, India

(2023-01 - 2026-05)

• Translated ambiguous operational requirements into AI solutions using OpenAI GPT-4o/GPT-4 and Google Gemini while balancing reliability, security, latency, and cost.
• Architected RAG and semantic-retrieval pipelines with FAISS, Pinecone, ChromaDB, embeddings, and hybrid retrieval across large unstructured-document collections.
• Built LLM evaluation pipelines across 10,000+ test cases measuring hallucination behavior, retrieval precision/recall, faithfulness, latency, and release regressions.
• Designed agentic and multi-agent workflows with LangChain, LlamaIndex, and LangGraph, including function calling, tool execution, contextual memory, and task decomposition.
• Implemented PII masking, RBAC, TLS 1.3, OAuth/JWT, and SOC2/GDPR-aligned privacy controls across cloud microservices and enterprise AI workflows.
• Engineered asynchronous FastAPI services with Docker, Redis, Celery, AWS, and GCP, supported by React/Next.js dashboards, CI/CD, and handover documentation.

Master of Business Administration - Business Analytics - Avantika University

Bachelor of Commerce - Computer Application - Future Vision College, Vikram University