DrawingAbhinav Rana
Open

Fig. 2 — Career

Where I’ve worked

Client work is listed as role, company, dates and scope. I don’t publish clients’ systems — the engineering I can describe in full is my own, and the older roles below.

Lead Software Developer

Oct 2024 — now

Forward-deployed lead across three enterprise clients.

  • IntuitApr 2026 — nowPlatform engineering and agentic systems.
  • VuclipMar 2025 — Mar 2026Go microservices for an OTT subscription platform.
  • New RelicNov 2024 — Feb 2025CI/CD and vulnerability-management automation.
  • Embedded with three enterprise clients in two years, owning delivery end to end inside each one's codebase, constraints and release process.
  • I don't publish clients' systems. Happy to talk through any of this directly.

Go · Python · AWS · PostgreSQL

Skills

What I reach for

Languages
Go · C++ · Python · Java · TypeScript · SQL
Agentic & LLM
Multi-agent orchestration · Graph-based agent architecture · State management · Context engineering · Agent reliability harnesses · Model Context Protocol (MCP) · Retrieval / RAG · Tool use & function calling · Multi-provider routing · Token budgeting
Backend
Microservices · Domain-driven design · REST · gRPC · ConnectRPC · GraphQL · Protocol Buffers · Envoy · Event-driven architecture
Data
PostgreSQL · SQLite · Schema design & migrations · Major-version database upgrades · Blue/green cutover
Cloud & Infra
AWS · Docker · CI/CD · GitHub Actions
Practices
Test-driven development · Code review · Security hardening · Structured logging · CI quality gates
Familiar with
LangChain / LangGraph · Anthropic Claude API · OpenAI API

Learning

How I got here

I learned to build agents by building them badly first, in the open. These are the steps, not the destination.

  • mini-perplexity → MINION ↗Started as a single research loop over three tools. Grew into a multi-tool agent with native tool use, an MCP server and a dashboard. The commit history is the lesson.
  • EAG V3 — Extensive AI for AgentsThe School of AI, 2026. Agent runtimes, live-graph memory, budget-aware planning, OpenTelemetry tracing for agents, and honest evaluation harnesses.

Next, not now

  • An eval harness with a rubric judge and a CI gate
  • OpenTelemetry GenAI tracing across an agent run
  • An MCP server written in Go

These are intentions, not claims. I haven’t built them yet.

Education

  • Master of Computer ScienceAmrita University, Coimbatore2016 — 2018 · CGPA 8.0
  • B.E. Electrical & ElectronicsVinayaka University, Salem2009 — 2013