Hi, I’m Akash Kumar.
I build agentic AI products.
👋 Senior software engineer with 8+ years building production AI systems, agentic workflows, and full-stack applications that ship.

A short version of the long story.
Tools I reach for.
Languages
3- TypeScript
- Python
- Go
Frameworks
12- React.js
- Next.js
- Tailwind CSS
- Node.js
- FastAPI
- Flask
- LangChain
- LangGraph
- CrewAI
- MCP
- AI Agents
- Temporal
Tools
- Cloud3
- AWS
- AWS AgentCore
- Serverless
- Data3
- PostgreSQL
- MongoDB
- Redis
- Infra2
- Docker
- Kubernetes
- Architecture3
- System Design
- Distributed Systems
- RAG
Recent projects.
Agentic AI platforms, full-stack applications, and production systems I’ve shipped.
Agent Builder Platform
Led end-to-end design of Agent Builder Platform with LangChain and Temporal. Built FastAPI-based Dynamic Agentic Registry for agent lifecycle management and React Flow Visual Workflow Builder for no-code multi-agent pipeline composition.
- LangChain
- Temporal
- FastAPI
- React Flow
- Python
- AWS
- Multi-Agent
Deep Research Agent
Architected production-grade Deep Research Agent using Temporal for durable workflow orchestration with multi-step parallel pipelines, dynamic branching, and stateful checkpointing. Enabled 100+ concurrent research sessions with fault-tolerant automatic retry.
- Temporal
- LangChain
- Python
- Distributed Systems
- Workflow Orchestration
CodeAct Agent
Built CodeAct Agent orchestrated via Temporal workflows where user intents are translated into executable code running in isolated Daytona sandboxes. Each activity provisions fresh workspace with full auditability.
- Temporal
- Daytona
- Python
- LangChain
- Code Execution
- Sandboxing
Agentic RAG System
Architected agentic RAG system with Temporal-orchestrated agent that iteratively refines retrieval queries. Performs dense vector search against pgvector knowledge base, evaluates context relevance, enabling accurate multi-hop reasoning.
- Temporal
- LangChain
- pgvector
- PostgreSQL
- RAG
- Vector Search
Multi-Agent Orchestration
Built managed multi-agent execution service on AWS AgentCore Runtime supporting CrewAI and LangGraph agents. Leveraged AgentCore Gateway as centralized MCP tool server with persistent cross-session state.
- AWS AgentCore
- CrewAI
- LangGraph
- MCP
- Multi-Agent Systems
Video Conferencing Platform
Engineered cross-platform video conferencing solution using Amazon Chime, React, Electron.js, and AWS Serverless. Designed 20+ serverless REST APIs with AWS Lambda for meeting orchestration and real-time collaboration.
- React
- Electron.js
- Amazon Chime
- AWS Lambda
- Node.js
- Serverless
League Management System
Architected and developed League Management System using React and Flask, enabling team registration, match scheduling, live score updates, player statistics, and automated notifications.
- React
- Flask
- Python
- PostgreSQL
- Real-time Updates
Experience
Where I’ve been.
- Feb 2025 - Present
Senior Software Engineer · Grid Dynamics
Bengaluru, India
Led end-to-end design of Agent Builder Platform with LangChain and Temporal. Built multi-agent orchestration on AWS AgentCore, architected Deep Research Agent with Temporal for durable workflow orchestration, developed CodeAct Agent with Daytona sandboxes, and implemented agentic RAG system with pgvector. - Jan 2021 - Jan 2025
Senior Software Engineer · Tech9
Pune, India (Remote)
Engineered cross-platform video conferencing solution using Amazon Chime, React, Electron.js, and AWS Serverless. Designed 20+ serverless REST APIs, resolved memory leak issues in Electron, and contributed to jewelry e-commerce platform redesign. Improved frontend performance through lazy loading and code splitting. - Aug 2018 - Jan 2021
Software Engineer · Ignite Solutions
Pune, India
Architected League Management System using React and Flask with team registration, match scheduling, and live score updates. Built secure payment service platform with Angular, integrating payment gateways and real-time transaction tracking. Optimized React application performance, reducing page load time by 40%. - Aug 2017 - Aug 2018
Full Stack Developer · Geek Informatic & Technology
Chandigarh, India
Developed video-centric job portal using MEAN stack with end-to-end features for video uploads and processing. Implemented video processing pipelines using FFmpeg. Configured and managed AWS infrastructure including ELB, Route 53, ensuring high availability and scalability. - Dec 2016 - Jun 2017
Software Developer Intern · Appnovation Technology Solutions
Hyderabad, India
Contributed to development of Smartwinnr, a gamified quiz platform for enterprise training using MEAN stack. Implemented real-time functionality using Socket.io, enabling instant updates and improving user engagement and collaboration.
Resume
My resume.
The full story of where I’ve worked and what I’ve shipped. Open the viewer to read on-screen, or grab a copy.
- 8+ yearsbuilding production AI agents and scalable web applications
- Agent Builder Platformdesigned and led at Grid Dynamics with LangChain + Temporal
- 100+ concurrent sessionson Temporal-orchestrated Deep Research pipelines
- 20+ serverless APIsshipped for a cross-platform Amazon Chime video product
Pulse from the repo.
Year at a glance for @akki907 — contributions, streaks, and the stack in heavy rotation.
2026 so far
consecutive days
best run in year
—
Stack snapshot
- TypeScript47%
- Python37%
- Go15%
Based on the languages I reach for most across recent work.
See the work on GitHub.
Side projects, experiments, and the occasional fix for tools I use.