Hey 😃 I'm Nir, an undergraduate majoring in Computer Science at Carnegie Mellon University. Here's some more about me:
- 💻 This semester I'm TAing 15-213: Introduction to Computer Systems at CMU. Very excited help students build a memory allocator, cache simulator, multithreaded web proxy, and more from the ground up in C! Last semester I TAd 15-122, CMU's imperative computing class. TAing is fun and fulfilling!
- 🔭 In 2026, I interned at Palantir working on infrastructure, with my proudest accomplishment being implementing optimizations that led to >10x cuts in memory and cost of a key service. The summer before, I worked as a SWE intern at Firebolt Inc, where I got to work with machine learning, distributed systems, and databases in a startup environment.
- 🏆 Hackathons:
- Recently built a autonomous bioreactor at the UC Berkeley AI Hackathon designed to help farm bacteriophages -- I enjoyed getting to combine hardware, software, and my robot learning knowledge into this project! See more here.
- Last year I won at CMU TartanHacks, CMU's largest hackathon (>1000 hackers and 279 submitted projects), with BallotGuide. It's a website designed to address low voter turnout with custom visualizations of how your local election ballot will actually affect your life. See this cool gif demo of the website's features.
- 📚 Recent papers I've enjoyed reading have focused on language modeling, robot learning, game theory/behavioral economics, genetic algorithms, and other fields in AI (hyperlinked one per category). Hit me up if you have any paper recommendations you think I'd enjoy!
- 📫 Reach me via my linkedin!
- 🥴 (Bonus: if you can't already tell, I love emojis)
Languages
Robot Caregiving and Human Interaction (RCHI) Lab @ CMU (Jan 2026 – Present)
- Under Prof. Zackory Erickson, working on robot learning for physical human-robot interaction.
- See the GenSimPHRI website for our recent work on text2sim2real, a framework for generating expert real-life policies entirely from short task descriptions -- no demonstrations or finetuning!
Assistive Sensing Systems (2020 – 2025)
- Invented and developed a device (patented: US 10830626 B2) that helps blind people prevent overpouring.
- Published research on using LiDAR ToF sensing for material level indication in DTIS 2025 (print pending).
Carnegie Mellon University | B.S. in Computer Science (GPA: 4.0/4.0).
Technical Courses: Completed
-
15-213: Introduction to Computer Systems
-
15-150: Functional Programming
-
15-122: Principles of Imperative Computation
-
15-151: Mathematical Foundations for Computer Science (discrete math)
-
21-241: Matrices and Linear Transformations
-
21-266: Vector Calculus using Matrix Algebra
-
21-270: Introduction to Mathematical Finance
Technical Courses: In Progress
- 07-280: AI & ML I
- 16-385: Computer Vision
- 15-210: Parallel and Sequential Data Structures and Algorithms
- 15-259: Probability and Computing
Software Engineer Intern – Palantir (Summer 2026)
- Deployed optimizations that cut a key service's memory usage and cost by >10x.
- Then, led a major refactor of a fleetwide service and mobilized two teammates alongside me.
Software Engineer Intern – Firebolt Inc. (Summer 2025)
- Trained an RNN model with AWS SageMaker to predict AWS EC2 capacity and customer demand for warmpool sizing, then deployed the entire data cleaning, featurization, and inference pipeline to a Go gRPC microservice.
- Developing a Go + gRPC tool that benchmarks data‑warehouse performance across cluster shapes, query acceleration settings, and caching strategies.
- Giving engineers a use-case-accurate perf-gate for new features and safeguarding Firebolt’s claim as the world’s fastest data warehouse.
Software Engineer Intern – SDF Labs → dbt Labs (Summer 2024)
- Led end‑to‑end development of the Reports cloud‑console feature (built with Svelte, TypeScript, TailwindCSS), used to track metadata across data workloads.
- Scraped & normalised SQL‑dialect metadata in Python to power cross‑dialect compilation.
Co‑founder & CEO – Extentek LLC (2018 ‑ 2025)
- Invented and patented (see US Patent US 10830626 B2) an over‑pour prevention device for blind users.
- Raised funding & spoke at global accessibility conferences (AHG, M‑Enabling, GeekWire).
Intern & Team Lead – SEAL Lab @ UW (Summer 2023)
- Managed a team of interns to write an NIH STTR R41 grant, bringing the lab’s software to millions.
| Project | Tech Stack | What It Does |
|---|---|---|
| 🧫 Phagentic |  ESP32/C++ · Python · PyTorch · BLE · MPC |
Built a physical bioreactor from scratch (ESP32, MOSFET-driven pumps, RGB sensor, 3D-printed rig) and trained a chemistry-informed gray-box ODE + MPC controller to autonomously drive its reaction. See more here. |
| 🪴 Verde | Flutter · Supabase |
Real-time map for community-driven environmental action. Scrapes and clusters issues, generates events, and gamifies eco-impact. Built in under 24 hours for HackMIT 2025. |
| 🔍 Truely | TypeScript · Svelte |
Annotates misinformation & deepfakes in-page via multi-agent LLMs and live web search. |
| ⚙️ Mini Prop Shop | Go · PyTorch · K8s |
Replays 1.2 B BTC ticks, serves sub-20 ms trading signals, and nightly fine-tunes an LSTM model. |
| 📈 Arduino Stock Ticker | C++ · Finnhub API |
Displays live price, symbol, %-change on a physical ticker—refreshed every minute. |
| 🧩 Canvas Submission Orchestrator | Python · Canvas API · Teams API |
Auto-assigns student submissions to TAs based on preferences, pinging assignments via Teams. |
| 🧦 Sock Match iOS | Swift · Create ML |
iOS app that lets blind users match socks with camera; on-device ML for real-time accessibility. |