AI Engineer — Agent Builder
Building agent systems
that work in production.
Retrieval Quality Baseline Sprint— $950, 5 business days
RAG pipelines and multi-agent workflows at Infradock.ai — with a track record across real-time systems and healthcare APIs.
Agent run
Retrieval workflow 042
Retrieval P@5
95.4%
p95 latency
740ms
Groundedness
0.91
Featured Work
Selected systems.
Direct product surfaces, agent infrastructure, and tooling built to ship rather than just demo.
01 — Inbox Agent
ReasonFlow
Autonomous inbox AI agent that classifies emails, retrieves context, and drafts responses with human-in-the-loop fallback when confidence drops.
94% confidence / sub-second retrieval
02 — Developer Tooling
MCPHub
Postman for MCP servers. Test, debug, and discover MCP servers in the browser without building custom local harnesses first.
15+ pre-configured servers
03 — Mission Control
AgentMesh
MCP-native multi-agent orchestrator with a real-time dashboard for monitoring coordination, execution state, and agent communication.
3 agent templates / live orchestration
Experience
Experience
Shipped in production.
Three roles across AI infrastructure, real-time systems, and healthcare APIs — each with measurable impact under real usage.
Infradock.ai
AI Engineer Intern
Building multi-agent workflows and RAG pipelines for enterprise clients, with retrieval quality pushed through semantic chunking and cross-encoder reranking.
95%+ retrieval precision on internal eval set
Oldowan Innovations
Software Developer Intern
Built Django matching APIs and Celery/Redis notification infrastructure on Stream Chat webhooks, load-tested with Locust to 10,000+ events/day at sub-50ms p95 latency, with retry and failure handling for webhook errors and worker restarts.
10k+ events/day load-tested (Locust)
Yantram Medtech
Software Engineer Intern
Built healthcare microservices with JWT-based RBAC and reduced database latency through targeted indexing and API performance work.
25% database latency reduction
Systems &
Execution.
Transforming complex technical requirements into physical logic. Interactive agent graph—hover to track node flow.
Honors and awards
Honors & Awards
Recognition with actual signal.
Four competitive finishes: global, national, and track-level, with real build pressure and limited execution time.
NASA Space Apps Challenge
Winner · Top 1% global
AI-driven orbital mechanics and tracking solution built for high-signal decision support under hackathon constraints.
CodeSlayer, NIT Delhi
Finalist · Top 1% of 10,000+
Systems-heavy build focused on architecture quality and execution speed in a competitive national setting.
MumbaiHacks
Finalist · Finance Tech track
Finance-oriented hack built around fast reasoning loops and a practical product flow that reached the finals.
Smart India Hackathon
Finalist · Top 5% nationwide
Problem-first system selected nationally that balanced strict constraints, scalability, and practical usability.
Blog
Technical
writing.
On AI systems, agent infrastructure, and production engineering.
What's next?
LET'S BUILD
THE FUTURE.
© 2026 Naman Gupta. All rights reserved.
Engineered with ♥ in India