Skip to content
View HrishiKabra's full-sized avatar

Highlights

  • Pro

Block or report HrishiKabra

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.

Content in all repositories owned by your account will be closed.
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
HrishiKabra/README.md
Hrishi Kabra — view from the surface, about to descend

Building person-enrichment pipelines @ Louisa AI · Publishing at AAMAS 2026 · Pricing tournaments with Dixon-Coles · 80 dives logged, favorite site: Komodo

Email LinkedIn Portfolio Resume

🤿 0m · Surface Interval

Hey, I'm Hrishi — a junior at Tulane University double-majoring in Computer Science and Mathematics (minors in AI and Business, 3.95 GPA) on a full-tuition South Asian Scholar merit award, based in New Orleans. I like building intelligent systems end to end: from the math underneath to the deployed product on top. Pre-dive check complete.

🔎 Recruiting for Summer 2027 SWE and AI/ML internships.

🧭 Dive Log · Experience

  • 🛠️ Louisa AISoftware Engineer Intern, Embedded AI team · relationship-intelligence platform spun out of Goldman Sachs. Building production person-enrichment pipelines: open-web fallback enrichment, confidence-gated LLM identity judging, and trust-tier field reconciliation.
  • 🗳️ Tulane Data & AI Collective (TUDAI)Research Assistant · co-authored a paper accepted at AAMAS 2026 on optimal positional scoring rules in computational social choice. Currently exploring social-choice methods for retrieval in multi-agent RAG.
  • 🧑‍🏫 Tulane Computer ScienceTeaching Assistant, CMPS 3160 (Data Science) & CMPS 1500 (Intro CS) · weekly labs and office hours for 50+ students.
  • 🤖 Niyogin AIAI Intern · designed and deployed an NLP pipeline for automated document digitization at 97% precision.
  • 📊 ITI Growth OpportunitiesIntern · researched and presented GenAI applications across industries.

🪸 18m · The Reef

Where the activity is. Six builds I keep coming back to:

RAG system over 1,000+ FIA regulation and steward-decision PDFs that explains Formula 1 incidents with rule-grounded citations.

  • GenAI Ops experiments on chunking, embeddings, and context size lifted citation coverage 60% → 93%
  • Python · Gradio · LLM APIs · RAG

Dixon-Coles bivariate Poisson model fit via MLE on 15,494 international matches, priced by an LMSR automated market maker over 100K Monte Carlo tournament simulations.

  • 0.572 Brier score vs. 0.613 Elo baseline on a walk-forward Euro 2024 backtest
  • Python · SciPy · FastAPI · Next.js · DigitalOcean · Vercel

🗳️ Optimal Voting — AAMAS 2026

Research framework and Python package for computing optimal positional scoring rules using simulated annealing, gradient descent, and MIP across four social welfare functions.

  • Co-authored with Ben Armstrong and Nicholas Mattei · Live demo
  • Python · Django · MIP

Async prediction-market arbitrage scanner across Kalshi and Polymarket with fuzzy event matching (Jaccard + Levenshtein, >95% accuracy) and fee-adjusted EV execution in under 2s end to end.

  • Python · AsyncIO · SQLite · Docker

🐟 FishID

Full-stack web app that identifies fish species from images — Flask REST backend integrating Fishial + Groq AI, JWT/bcrypt auth, split deployment on Vercel and Render.

  • Python · Flask · Next.js · TypeScript · SQLite

F1 telemetry fingerprint tool visualizing throttle behavior across circuits using 360° resampled OpenF1 data.

  • Python · OpenF1 API

🔬 30m · The Research Station

  • 📄 Co-authored paper accepted at AAMAS 2026 (Conference on Autonomous Agents and Multiagent Systems) on optimal positional scoring rules — with Ben Armstrong and Nicholas Mattei
  • 📦 Shipped the research as a PyPI package: optimal-voting · live demo
  • 🧪 Current direction: social-choice methods for retrieval in multi-agent RAG
  • 🎓 Dean's List every semester · full-tuition South Asian Scholar (~$67K/yr) · Corporate Sponsorship Chair, Alpha Kappa Psi

Instrumentation on board:

Languages — Python · Java · C/C++ · JavaScript · TypeScript · SQL · R · Assembly · Bash AI/ML — PyTorch · RAG · LLM Agents · NLP (SpaCy) · Prompt Engineering · Fine-Tuning · MLOps · Evals Full Stack & Infra — React · Next.js · Flask · Django · FastAPI · AWS (S3, Glue, Athena) · Docker · Git · SQLite

⚓ 40m · The Wreck

Older builds, still worth the swim-through:

  • 🧭 Wikipedia Race — bidirectional BFS pathfinding game across Wikipedia
  • 🐋 Ocean Jukebox — 131 NOAA hydrophone recordings, browsable by ear
  • 🌊 ReefGuardian — serverless AWS pipeline over NOAA sea-surface temperatures, flagging coral-bleaching hotspots
  • 🎸 GigPilot — multi-agent live-music booking marketplace

🌑 60m · The Abyss

Live telemetry from the bottom. The two instrument readouts regenerate every night from my actual GitHub activity:

Dive computer — live GitHub stats as dive readouts Dive forecast — Brier-scored probabilities about my own activity

An eel eating my contribution graph 3D contribution terrain
⬇️ descend further — full telemetry
Isocalendar and language metrics

GitHub stats Top languages

🫧 5m · Safety Stop

Three minutes before surfacing. Off the keyboard I'm a PADI Master Scuba Diver and underwater videographer — 80 dives logged, with Komodo, Indonesia as my favorite site — and I think a lot about where intelligent systems and ocean conservation can meet. Also: Formula 1 weekends, a long-running Wordle streak, and the occasional ruthless Monopoly tournament.

— surfacing · thanks for diving along 🫧 · say hi

Pinned Loading

  1. reefscan reefscan Public

    End-to-end coral reef health analysis — SAM2 + DINOv2 + conformal prediction, FastAPI + Next.js, deployed on free tier

    Python

  2. pnyx pnyx Public

    When do prediction markets beat static pools? A controlled stress test of LLM belief aggregation against a known Bayes-optimal posterior. LMSR markets, quality-matched teams, adversaries, and a cap…

    Python

  3. voting-mcp voting-mcp Public

    MCP server exposing principled social-choice aggregation rules + a benchmark measuring their accuracy vs majority vote over an LLM ensemble

    Python

  4. fishidapp fishidapp Public

    TypeScript

  5. AI_Engineering_Project AI_Engineering_Project Public

    Jupyter Notebook

  6. wc-prediction-market wc-prediction-market Public

    Python