Architected a reproducible ReAct-style research agent integrating MCP HTTP tool suites, dynamic web retrieval, typed schema validation, and persistent memory logging.
MIHIR SUHANDA
01 Executive Profile
AI Engineer specializing in production machine learning, autonomous agentic workflows (Model Context Protocol & PydanticAI), and high-throughput RAG architectures. Backed by AI engineering specialization at Saras AI Institute (Grade A) and 2+ years of mission-critical engineering execution at Toyo Engineering. Bridges high-reliability industrial systems design, automated data pipelines, and scalable cloud deployments (AWS, Docker, FastAPI).
02 Flagship AI & Agentic Systems
Built an enterprise-grade inference API achieving 95%+ accuracy with real-time risk scoring, feature engineering pipelines, and Docker deployment.
Engineered a dense vector embedding pipeline with chunking optimization and semantic retrieval, enabling natural-language querying over complex technical documents.
Deployed an event-driven AWS serverless architecture delivering personalized portfolio allocations, integrating ML risk scoring models and AWS Lambda.
Designed a CLI music manager with role-based access control and database-backed CRUD features, demonstrating robust backend architecture.
Interactive web application for intelligent travel itinerary planning, geolocation filtering, and real-time destination route management.
03 Professional Experience
- Spearheaded engineering execution on mega-scale EPC packages including the VGO-HDT unit and Ammonia Storage Tanks (2 × 20,000 MT AST), ensuring strict compliance and safety standards.
- Streamlined instrumentation & automation workflows, reducing inter-team rework by 18% and accelerating cross-functional alignment across 5+ stakeholder groups.
- Enforced strict fail-safe reliability standards and data integrity across complex control loops in high-pressure project delivery environments.