Pawan Kumar

Software engineer · Systems design & applied AI

System design and applied AI.

I design and build distributed systems, APIs, and applications that use language models. Recent work includes a device orchestration platform for hardware validation, telecom log and KPI analysis with OpenSearch and LLMs, and tools for tracking test reliability. I focus on clear interfaces, observable behavior, and dependable operation.

System designDistributed platformsLLM applications
A few selected projects ↓ Get in touch ↗
From design to productionPlan → build → improve
A steadily rising path through four engineering steps Understand → Design → Build → Operate 01 / PROBLEM 02 / DESIGN 03 / BUILD 04 / OPERATE Understand Define Integrate Strengthen users · data · constraints service boundaries · API contracts services · APIs · language models tests · monitoring · recovery
Turning requirements into working softwareUnderstand → Design → Build → Operate
~15%fewer API failures after production issue fixes
~40%more regression coverage through simulator improvements

01 / Experience

Industry experience

Backend platform development, system reliability, and production engineering.

Mar 2026 to Present

Client engagement

Zebra Technologies

Software Engineer · Platform Engineering

Designed and built a device orchestration platform for hardware validation. It uses a stateful LangGraph agent, typed MCP tools, and durable job recovery.

  • Led architecture and development of the hardware validation platform, including its execution engine and tool integrations.
  • Built a stateful LangGraph agent for device tool routing and concurrent workflows using Python asyncio.
  • Implemented an MCP server with typed tools for device control, diagnostics, log retrieval, and Git operations.
  • Added retries, safeguards, logging, and telemetry to LiteLLM tool calls. Persisted jobs in SQLite through aiosqlite for recovery.

Tools · LangGraph · LiteLLM · MCP · Python asyncio · SQLite · aiosqlite

Aug 2024 to Present

Employer

Telaverge Communications

Software Engineer · Research & Development

Built backend APIs and reliability tools, and contributed to telecom log and KPI analysis using OpenSearch retrieval and response caching.

~15%Fewer API failuresAfter production issue fixes
~40%More regression coverageThrough simulator improvements
  • Built REST APIs and a Python/MongoDB marketplace for third-party application onboarding and configuration.
  • Investigated API timeouts, serialization failures, and downstream bottlenecks.
  • Extended an HTTP/2 simulator with delay injection and dynamic plugins for automated regression testing.
  • Contributed to log and KPI analysis using an LLM. Built OpenSearch retrieval and context assembly pipelines with response caching for common queries.

Tools · Python · MongoDB · OpenSearch · LLM retrieval · HTTP/2

02 / Independent work

Selected projects

Selected projects across system design, developer tools, and applied AI.

GitHub Live demo
GitHub Live demo

Backend engineering · LLM integrations

I build backend platforms and LLM features.

I work on service design, data pipelines, and language model integrations. Recent work includes a device orchestration platform with a stateful LangGraph agent, OpenSearch retrieval for telecom analysis, and APIs and test infrastructure. I aim for software that is observable, recoverable, and straightforward to maintain.

Technologies and areas of focus Python · System design · Distributed systems · FastAPI · Django REST Framework · PostgreSQL · MongoDB · Redis · OpenSearch · LangGraph · LiteLLM · MCP · RAG · Docker

Contact · Bengaluru, India

I’d welcome a conversation.

Email GitHub LinkedIn LeetCode