Hi, I’m Arth.
I live in Austin where I work as a machine learning engineer at Tesla focused on developing algorithms and scaling infrastructure for computer vision based defect detection systems.
Before Tesla I graduated from UC Berkeley where I majored in Electrical Engineering and Computer Sciences. While there I:
- Wrote a creative column for one of the oldest college newspapers in the country.
- Published research on improving the safety of semiautonomous vehicles.
- Learned how to count cards from Vegas veterans.
- Investigated human trafficking in New York City.
- Pitched a medical drone delivery network to an Indian Chief Minister.
- Wrote mapping software for a fleet of autonomous cars.
- Analyzed data for a non-profit working in a Jordanian refugee camp.
- Built several ML + mobile healthcare tools for UCSF.
- Led one of Cal’s oldest consulting groups.
- Jumped out of a plane.
I’m interested in statistical learning, high performance computing, back-end infrastructure, reliability, computer vision, and optimization.
I also maintain a newsletter called Divergence which you can learn more about below.
Training Data
Experience
Tesla; ML Eng
Net Impact Berkeley; President
VeHICal; Research
DeepMap; Software Eng
VeHICal; Research
RISELab; Research
UCSF; Research
Berkeley SCET; Research/Eng
CS61A; Student
Contribution
Developing algorithms and scaling infrastructure for hybrid deep learning and classical computer vision based defect detection.
Lead a team of 37 people at UC Berkeley’s third oldest consulting group working on five social impact driven projects and helped secure Gold Status for Cal’s undergraduate chapter.
Published techniques for scoring the credibility of perception modules in semi-autonomous vehicles.
Implemented end-to-end trainable multi-modal architecture for LiDAR-Camera Fusion to develop better maps for autonomous fleets.
Used generative AI to model driver attention in semi-autonomous vehicles and improve control handoff from the vehicle to the driver.
Worked on reinforcement learning algorithms for high dimensional path planning.
Developed semantic segmentation models for predicting cardiac disease from echocardiograms.
Designed and pitched a proposal to the Chief Minister of Meghalaya for the first ever drone powered medical supply network for rural communities in his state.
Wrote my first line of code.
When
Late 2022
– Now
Early 2022
Late 2021
Mid 2021
Early 2021
Early 2021
Early 2020
Late 2019
Early 2019
Divergence
A newsletter about the hidden connections between some of the most powerful ideas across disciplines of human knowledge and their implications on the world.
The writing is split between Substack and Notion (for when the math gets heavy).
Corrections are always welcome.
Arth Vidyarthi 