Manas Choudhary

Intern

Palwal, Haryana, India0 mo experience

AI EnabledAI ML Practitioner

Key Highlights

Architected enterprise-grade AI agent infrastructure adopted by 70% of teams.
Achieved 98% accuracy in Figma-to-React AI pipeline.
Scaled no-code AI agent platform to 100+ monthly active users.

Stackforce AI infers this person is a SaaS-focused AI Engineer with expertise in multi-agent systems and web development.

Contact

Skills

Core Skills

Ai AgentsLarge Language Models (llm)System Design & ArchitectureWeb DevelopmentDeep LearningRoboticsUi/ux

Other Skills

Multi-agent SystemsAgent-based ModelingMulti-Agent AI SystemsHarness engineeringmcpLangChainLangGraphDockerKubernetesWeb Content Accessibility Guidelines (WCAG)Figma (Software)TypeScriptFigmaTailwind CSSGitHub Actions

About

AI Systems Engineer building production agent infrastructure, LLM bias mitigation systems, and open-source governance tools. At Atlan (Series C, $750M — Gartner Leader in Data & AI Governance), I architect enterprise AI agent systems using MCP, LangChain, and LangGraph — adopted by 70% of engineering teams. At IIT Roorkee, I'm building an agentic red-teaming framework that detected 40+ unique bias patterns beyond standard benchmarks. Previously: shipped a Figma-to-React AI pipeline at Samsung R&D achieving 98% accuracy across 50+ components, reducing manual UI effort by 70-80%. Founded Sigmoyd, a no-code AI agent platform scaled to 100+ MAU. Contributed 15+ merged PRs to Google DeepMind's Concordia framework (1.2k+ stars). Core stack: Python, TypeScript, Rust, LangChain/LangGraph, MCP, Docker, Kubernetes, vLLM, Redis, PostgreSQL, FastAPI. Open source: govrix.dev | github.com/manaspros Portfolio: manasdev.codes Email: manas24102@iiitnr.edu.in

Experience

0 mo

Total Experience

Average Tenure

Current Experience

Indian institute of technology, roorkee

Research Intern

Feb 2026 – Present · 4 mos · Remote

Cant tell until done

Multi-agent SystemsAgent-based ModelingMulti-Agent AI Systems

Atlan

AI Engineer

Jan 2026 – Apr 2026 · 3 mos · 3 Coleman Street, #03-24, Singapore · Remote

Architected enterprise-grade AI agent infrastructur using MCP (Model Context Protocol) with LangChain and LangGraph, adopted by 70% of engineering team; orchestrated stateful context management with Redi-backed caching across multiple LLM providers, deploed on Docker and Kubernetes.
Eliminated race conditons in async tool-call pipelines, improving system utime from 75% to 95% and cutting mean time to recovery by 60%; instrumented end-to-end observability with Langfuse and LangSmith.

Harness engineeringmcpLarge Language Models (LLM)AI AgentsMulti-Agent AI Systems

Google deepmind

Open Source Contributor

Dec 2025 – Feb 2026 · 2 mos · Remote

Delivered production-critical contributions (15+ meged PRs) to Concordia, DeepMind's generative agent smulation framework for AI safety research (1.2k+ Gitub stars).
Fixed crash-level failures in core simultion engine, migrated structured logging across 26 fles, and improved validation and robustness of agentexecution and reasoning modules.
GitHub: https://gihub.com/google-deepmind/concordia

Samsung r&d institute india - bangalore

R&D Intern – AI & Design Automation

Oct 2025 – Apr 2026 · 6 mos · Remote

Created a bidirectional Figma-to-React TypeScript sync sstem with Tailwind CSS, achieving 98% accuracy acros 50+ UI components with GitHub Actions CI/CD and Verel deployment.
Reduced manual UI development effort y 70-80% via an LLM-driven extraction pipeline (Pythn, Node.js) with WCAG accessibility compliance and RST API integration, deployed across Samsung product eams.

Web Content Accessibility Guidelines (WCAG)Figma (Software)TypeScriptAI AgentsSystem Design & Architecture

Indian institute of technology jodhpur

Summer Research Intern – Robotics, AI & Computer Vision

May 2025 – Jul 2025 · 2 mos · Jodhpur, Rajasthan, India · On-site

Implemented 4 ROS-based robotic systems (Python, Tensorlow, OpenCV, CNNs) on Linux with real-time computer ision pipelines achieving 92% detection accuracy and under 100ms pose estimation latency.

Deep LearningRobot Operating System (ROS)GazebosField ResearchRobotics

Sigmoyd

Software Engineer

Mar 2025 – Oct 2025 · 7 mos · Raipur, Chhattisgarh, India · Hybrid

Scaled a no-code AI agent platform (Next.js, Express.js, MongoDB Firebase) to 100+ monthly active users on AWS and GCP, enabling agent deployment across Gmal, Slack, Notion, and WhatsApp with 20+ MCP-powered rchestration templates and event-driven triggers.
Sipped a persistent-memory voice agent using Mem0 andWebSocket streaming for cross-session context, containerized with Docker and compatibl with any MCP connector for cross-platform workflow utomation.

Node.jscomposioReact jsFastAPIAgentsWebSocket+4

Tribal research and training institute raipur

Web Development Intern — Tribal Freedom Fighter Museum Website

Feb 2025 – May 2025 · 3 mos · Remote

Developed and currently deploying the official website for the Tribal Freedom Fighter Museum, Naya Raipur under the guidance of faculty from IIIT-NR.
🔹 Built using ReactJS, Three.js, Framer Motion, and Styled Components
🔹 Features include: virtual tour section, multilingual support, booking module, artifact showcase, and animated UI
🔹 Integrated rich tribal history and heritage with modern responsive design
🔹 Over 300 hours invested in development and optimization

ReactJSJavaScriptNext.jsTailwind CSSFramer MotionUI/UX+1