Balanced AI engineer profile — agent orchestration, knowledge graphs, team leadership.
AI Engineer and Tech Lead who ships production systems fast. Built a 49-tool MCP platform, led a 9-person AI team, and shipped 4 production AI systems in one month. Deep expertise in LLM agents, RAG, FastAPI, and the Anthropic/Claude ecosystem.
- Built and lead a 9-person AI team from scratch; introduced Git, CI/CD, code review, and sprint processes to the division
- Architected multi-tenant agent orchestration platform with 4 agent types (SQL, voice, email, workflow), 16+ tools, and DAG workflow execution serving a 650+ vessel fleet
- Created 49-tool MCP sprint command center for AI-driven development coordination across 9 developers, adopted by director within 24 hours of demo
- Built AI-native maritime ERP (FastAPI + React/TypeScript), contributing 49% of commits across 66 tasks and 92 dependencies
- Developed 3-phase QA automation system (viewpoint generation at 80-90% accuracy, test case generation, email automation) targeting 20% QA load reduction
- Built enterprise document search with hybrid Reciprocal Rank Fusion scoring, deployed to 5+ clients
- Shipped financial SQL agent with 11-step ETL pipeline and 88%+ query accuracy over millions of rows
- Built AI agents using LLMs (OpenAI, Anthropic, Hugging Face) with RAG and tool integrations
- Led Django to FastAPI migration; introduced Git version control to a team with none
- Built real-time communication infrastructure using Django Channels WebSockets
- Deployed production systems on Azure with CI/CD pipelines
Full-stack AI platform (FastAPI + PostgreSQL + Jinja2). 49 MCP tools, 27+ database tables, knowledge graph (680+ nodes, 1203+ edges), 12 entity types, RBAC auth, WebSocket real-time sessions, CI/CD on Railway. Built and deployed solo.
Sanctions screening with OFAC/EU/UN lists, AI document analysis, entity intelligence dossiers, audit trail. Full app framework with dual-mode auth and graph-native entities.
Multi-tenant platform with SQL/voice/email/workflow agents, TTL+LRU caching, Prefect DAG executor, WebSocket voice agents, MS Graph integration. Led team of 3.
Sprint planning, task dependencies, knowledge graph, agent check-in/checkout, file conflict detection. Coordinated 9 developers in parallel via MCP protocol.
Grew AI division from 1 to 9 engineers. Trained 15 interns. Coordinated with 6+ cross-functional teams.
Shipped 4 production systems in 1 month. Built Cortex (49 MCP tools, 27 tables) and Comply (B2B SaaS) solo. Direct delivery to enterprise maritime client (650+ vessel fleet).
Pitch Deck
12 slides covering Cortex architecture, team leadership experience, and technical vision.
Download Pitch DeckWant to talk instead of read?
aroraryan826@gmail.com →