About Me
About Me
Hey! I’m Avi — a graduate research assistant at CMU AirLab and a Master’s student in Computer and Civil Engineering at Carnegie Mellon University, graduating in December 2026. I build autonomous systems that connect perception, planning, and control, with a focus on getting algorithms working in simulation and on real hardware.
My work spans drone autonomy, adaptive flight control, computer vision, and machine learning. Across research, industry, and founding a company, I’ve built everything from embedded C++ controller frameworks to cloud-based multi-agent simulation platforms.
Aurora Flight Sciences
From May to August 2026, I was an Autonomy Intern at Aurora Flight Sciences, a Boeing company, working on the DARPA LINC-funded FALCON program. I extended adaptive control capabilities to quadrotors and built a swappable PX4 rate-controller architecture supporting PID, RSLQR, OBLTR, and AAOBLTR. The tuned system achieved 0.06–0.07 rad/s roll/pitch tracking error under disturbance, and I led a live quadcopter demonstration for Boeing and DARPA stakeholders.
I also built a TensorRT-accelerated HLoc localization pipeline with YOLOv11-n, eliminating reliance on AprilTag markers and delivering 3× more inliers and a 20% improvement in pose-estimate accuracy. For obstacle avoidance, I formulated a closed-form control barrier function safety filter, validated in simulation.
Research at CMU AirLab
At AirLab, I’m developing informative semantic path planning and 3D mapping for Shimizu’s construction-site surveys. This work combines aerial sensing, object recognition, coverage planning, and excavation-surface analysis, enabling volumetric differencing accurate to ±5 cm. I’ve also implemented multi-agent capability in AirStack, the lab’s open-source drone simulation platform built on NVIDIA Isaac Sim.
In our September 2026 HighBay tests, two drones flew separate survey missions and each observed every required coverage target on one battery. Their maps were then merged into a shared site representation.
Another highlight: I helped design, build, and fly CMU’s largest university-built drone, a roughly 140 lb dodecacopter with a custom carbon-fiber airframe and approximately 600 lb of thrust. The project received GoAERO Stage 1 and University Innovation awards.
See the Shimizu project and flight videos
Building Systems & Leading Teams
From May 2022 to August 2026, I was Founder and Technical Lead of Codex Labs LLC, where I led development of SWARMS, a cloud-based drone simulation platform acquired by MIT Lincoln Laboratory. Alongside engineering, I handled customer outreach, product demonstrations, and platform strategy.
Across three summer internships at Saab (2023–2025), I worked on federated learning, sensor fusion, synthetic training data, and real-time vision systems. I trained a drone detector on 10,000 synthetic images with 89% precision on real-world data, built an asynchronous C++ inference pipeline in Unreal Engine, and automated image annotation for model training.
At Purdue, I led drone SWARMS research in SLAM and semantic obstacle avoidance using LiDAR and cameras, presenting the work four times at the Undergraduate Research Expo.
Beyond Flight
My graduate projects also explore diffusion models, reinforcement learning, and learning-based control: a transformer-based image-editing backbone, PPO and SAC agents for Franka Kitchen manipulation, and a MuJoCo quadrotor simulator supporting learning-augmented MPC and neural Lyapunov certification. I placed in the top 10% of CMU’s deep-learning Kaggle competitions.
Robotics has been part of my life since VEX: I was a four-time Worlds qualifier, World Quarter Finalist, and mentor, with a team ranked in the world’s top 25. Away from the lab and coursework, you’ll usually find me on the basketball court.
Education
| Degree | Institution | Completion |
|---|---|---|
| M.S. Computer and Civil Engineering | Carnegie Mellon University | Expected December 2026 |
| B.S. Computer Engineering | Purdue University | May 2025 |
At CMU, I’m a fully funded graduate research assistant with the Robotics Institute (Courtesy Student status), with a 3.8/4.0 GPA.
Experience
| Role | Organization | Period |
|---|---|---|
| Graduate Research Assistant | CMU AirLab | Aug 2025 – Present |
| Autonomy Intern | Aurora Flight Sciences (Boeing) | May–Aug 2026 |
| Founder and Technical Lead | Codex Labs LLC | May 2022 – Aug 2026 |
| Machine Learning Software Engineering Intern | Saab | Summers 2023, 2024 & 2025 |
| Lead Drone SWARMS Researcher | Purdue University | Aug 2021 – May 2025 |