Building Devin, the world's first AI Software Engineer.

Contributing to the development of a generalized framework for parallel algorithms for spatial grid-based clustering. Developing sequential baselines for the Grid-based Density-Isoline Clustering (GDILC) algorithm to verify correctness and measure performance overheads and work efficiency of parallel implementations. Optimizing parallel implementations of clustering algorithms to improve performance in practical, limited-scale use cases. Exploring the design of novel parallel grid-based clustering algorithms by combining key ideas from existing algorithms.

Co-designing algorithms and hardware to design energy efficient solutions to localization and mapping tasks

Benchmarking algorithms for SLAM and multi-view stereo tasks to analyze memory-efficiency and throughput

Teaching topics in computer architecture, digital logic, operating systems to students during office hours.

Participated in mock trading to obtain familiarity with how the market operates at the level of individual orders. Competed in automated trading strategy and design competition and in an complex zero-sum game.