Skip to content
View prasad-yashdeep's full-sized avatar

Highlights

  • Pro

Organizations

@Active-Learning-and-Teaching

Block or report prasad-yashdeep

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
prasad-yashdeep/README.md
⌜                                                            ⌝

                   SYSTEM ━━ NOTIFICATION

          you have encountered a  HUNTER.

⌞                                                            ⌟

〈 Yashdeep Prasad 〉 ⚔

📍 Brooklyn, NY · 🧠 AI Agentic Developer · 🎓 MS Computer Engineering @ NYU (class of '26)

Python TypeScript Swift PyTorch FastAPI React Three.js Claude OpenAI AWS macOS

Deep in agentic-hackathon mode — shipping voice-controlled multi-agent systems, self-healing infra, and research that blurs the line between world-models and task-agents. I believe every shell should talk back and every cron should have a rank.

🔨 currently — learning-based GPU optimization for AI/DL workloads at DeepLM · JEPA world-models at NYU CDS · building my own agent runtime on weekends.


〈 ⌬ 〉 Current Projects

the System has granted you access to these dungeons.

  • 🎆 HunterOS — Voice-controlled Solo-Leveling command bridge for OpenClaw agent fleets. r3f hologram sigil, Whisper → intent → ElevenLabs, live Hunter cards with rank formulas, destructive-intent approval gate. TypeScript · FastAPI · Three.js
  • 🔧 AgentOps — Self-healing DevOps agent with human-in-the-loop approvals. Real fault injection → diagnosis → auto-fix with role-based clearance. Built for Iterate × CBS AI Hackathon. 4 sponsor stacks live. Python · Blaxel · Anthropic · ElevenLabs
  • 🩺 ML Sports Injury Prediction — Classifier predicting athlete injury risk with interpretable feature importance. ⭐ 11 · most-starred repo. Jupyter · scikit-learn · PyTorch
  • 🦞 OpenClaw (contributor) — Personal AI assistant that runs anywhere. Multi-agent, any OS. I contribute patches upstream + ship downstream skills. TypeScript · Node · multi-provider LLM
  • 🧮 gemma-function-calling — Training & evaluating Gemma for structured tool calls. ⭐ 3. Python · PyTorch · Transformers
  • 🛰️ NLP_MCP_AGENTS — Multi-agent NLP pipeline over the Model Context Protocol. Python · MCP
  • 💼 JobSearchAgent — Scan → score → outreach-draft pipeline. Browser automation + LLM ranking + Google Sheets as the control plane. Python · Playwright
  • 🧭 GSoC_MCP_TOOL25 — Google Summer of Code 2025 · DeepMind · agentic AI framework tooling. JavaScript · MCP
  • 🌐 prasad-yashdeep.github.io — Personal research site: publications, news, JEPA & agentic-AI notes. HTML

Research / Coursework

  • 🧠 llm_research — LLM research scratchpad: tabular feature stability, structured-data modeling, prompt-evolution experiments. Jupyter
  • 🔎 GemmaGSOC — Gemma toolformer experiments under Google Summer of Code.
  • 🧬 IQB_AFP — Antifungal peptide sequence classifier (undergrad bioinformatics). Python
  • 🗺 GIS-Analysis-with-GeoPandas — Spatial analysis coursework. Jupyter
  • 🌱 PlantsVsZombies-SY — Undergrad Java project (for the meme). Java

〈 ⚔ 〉 Stack I reach for

 LANGUAGES  ▌ Python · TypeScript · Swift · SQL · C++
 AI / ML    ▌ PyTorch · scikit-learn · LangChain · DSPy · MCP
             Anthropic · OpenAI · xAI · ElevenLabs
 BACKEND    ▌ FastAPI · Node · Postgres · Redis · WebSockets
 FRONTEND   ▌ React · Next.js · Vite · Three.js · r3f · Tailwind
 INFRA      ▌ AWS CDK · Docker · Cloudflare · GitHub Actions
 AGENTIC    ▌ OpenClaw · Claude Code · Cursor · Remotion · Playwright

〈 🌀 〉 What I'm doing now

  • 🪄 Building agent runtimes — voice-first dashboards for multi-agent orchestration (HunterOS)
  • 🧠 World-model research — JEPA self-supervised learning for embodied AI at NYU CDS
  • 💡 Shipping at ludicrous speed — hackathons, 1-day MVPs, commit → deploy → demo loops
  • 🎙️ Teaching agents to talk back — Whisper + ElevenLabs as a first-class UI layer
  • 📚 Writing on agentic workflows, self-healing systems, and the weird physics of multi-model orchestration (site coming soon)

〈 📜 〉 Recognition

  • 🎓 Google Summer of Code 2025 — DeepMind · agentic AI frameworks
  • 🧑‍🏫 Teaching Assistant — Intro to Deep Learning · NYU Courant
  • 🔬 Research student — NYU Center for Data Science · JEPA world-models
  • 💼 AI Agentic DeveloperDeepLM · GPU workload optimization
  • 🎒 BTech CSIIIT Delhi (prior to NYU)

〈 📡 〉 Connect

Website LinkedIn Scholar Email GitHub


⌜                                                            ⌝

                      " A R I S E . "

⌞                                                            ⌟

building agents that ship · one dungeon at a time

Pinned Loading

  1. gemma-function-calling gemma-function-calling Public

    Python 3

  2. dino_wm dino_wm Public

    Forked from gaoyuezhou/dino_wm

    Python

  3. NLP_MCP_AGENTS NLP_MCP_AGENTS Public

    Python 1

  4. tmforum-innovation-hub/ReasoningAgent tmforum-innovation-hub/ReasoningAgent Public

    Jupyter Notebook

  5. Active-Learning-and-Teaching/ALT Active-Learning-and-Teaching/ALT Public

    An open-source mobile application to facilitate active learning and teaching.

    TypeScript 22 9

  6. Bhumika-Shetty/AdFame Bhumika-Shetty/AdFame Public

    ML-ops Final Project

    Jupyter Notebook 2