Skip to content
View suhaasteja's full-sized avatar
13 followers · 44 following

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull SharkAchievement: YOLO

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull SharkAchievement: YOLO

Block or report suhaasteja

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
suhaasteja/README.md

Suhaas Teja — LLM Agent & AI Infrastructure Engineer

Building production AI systems at the intersection of agent evaluation, LLM infra, and applied ML.
Based in San Jose, CA · Open to collaboration on AI tooling, agent frameworks, and dev infrastructure.

What I Build

I design and ship AI agent systems — from multi-tier memory architectures to adversarial evaluation pipelines. My work focuses on making LLM agents more reliable, observable, and production-ready.

  • Agent Infrastructure — SageMem: GPU-inspired L1/L2/L3/DRAM memory hierarchy for multi-agent systems using Redis + pgvector
  • Agent Observability — playback: Session log visualizer that turns raw JSONL agent traces into step-by-step timelines
  • Adversarial Evaluation — gauntlet: CLI tool spawning 6 AI personas to red-team PRDs and codebases for weaknesses
  • Developer Tooling — claude-deck: Claude Code statusline showing context %, model, API usage, and Spotify track
  • Applied AI — NFL-contract-coach: AI platform reviewing sports contracts for predatory clauses across 10 risk categories
  • Embodied AI — reachy-rx: Pharmacist robot on Reachy-mini (3rd place, Seeed Studio hackathon)

Tech Stack

Languages: Python · TypeScript · JavaScript · Shell · C++
AI/ML: LangChain · LlamaIndex · DSPy · PyTorch · Ray · FastAPI · MCP · OpenAI · Anthropic · Gemini
Infra: Redis · PostgreSQL + pgvector · AWS (ECS, S3, CloudFront, DynamoDB) · Docker
Frontend: React · Next.js · Tailwind CSS · Convex · Liveblocks

Currently Working On

→ GenAI_notebooks — curated notebooks on LLM fine-tuning, DSPy, NVIDIA NeMo, and agent evaluation

Get In Touch

📫 suhaastejav@gmail.com
🔗 Portfolio · LinkedIn · Twitter/X

Pinned Loading

  1. playback playback Public

    Visualizes LLM agent session logs as step-by-step timelines — user ask, tool call, outcome

    JavaScript 2 1

  2. Flowdeck Flowdeck Public

    Browser-based HMI for simulated UR5 robot arm — real-time control and visualization interface

    JavaScript

  3. reachy-rx reachy-rx Public

    [SeeedStudio Embodied AI hack 3rd place] An embodied AI pharmacist robot that watches, reminds, and cares. Built for Reachy-mini robot.

    Python 1

  4. Lekha Lekha Public

    Google Docs clone with inline AI commands — generate, web search, CSV visualization, and Mermaid diagrams

    TypeScript

  5. gauntlet gauntlet Public

    Bash CLI spawning 6 persona-based AI agents to adversarially red-team PRDs and codebases

    Shell

  6. synapse-AI-ops synapse-AI-ops Public

    AI video calling agent for on-field data center technicians — multi-agent GPU observability platform

    Python