Kumar Chandra Email & Phone Number

California, United States

California, United States

Orlando, Florida, United States

Pursuing part time role for '24 - '25 Academic Year

Orlando, Florida, United States

Worked directly with lead researcher at Beep, deploying autonomous shuttles for transit agencies across the USApplied Discrete Event Simulation to model the individual passengers and shuttles in transit systems with 50,000+ daily passengers and 30+ routes. Made proprietary algorithms to determine the optimal amount of shuttles along routes to meet specified performance metrics, allowing transit agencies to estimate the cost of using Beep Shuttles Developed algorithms to change stop durations of autonomous shuttles in real time to prevent delays during severe congestion. Prevented >150 hours of traffic delays per day across all shuttles in simulations of a large transit systemAuthored “Dwell Time and Fleet Size Optimization for Autonomous Transit Operations” (White Paper at Beep)Created a React App using deck.gl to animate simulated events in a transit system at an individual passenger level. Can change parameters for specific agencies and re-optimize and re-visualize in real-time.

Cofounded startup working with Stanford Statistics Faculty to develop Machine Learning models with Tensorflow to predict the outcome of sports games, over $5000 profit within 2 months of operations

Stanford, California, United States

Improved upon approach to design optimal zones for SF schools by reformulating Mixed Integer Program as Constraint Program, leading to >84x speed increase from lab’s previous approach (Published in ACM Econ + CS)“SFUSD School Zone Design using Constraint Programming”, “Integer vs. Boolean Variables for School Rezoning”Developing a heuristic approach to generate zones for larger districting using Monte Carlo Markov Chain models

Boston, Massachusetts, United States

Worked with NC School district to produce rezoning algorithms to reduce school income segregation using Constraint Programming with Google CP - SAT, NumPy, Pandas, and NetworkX“Race or SES? Estimating the impacts of different policy objectives on K-12 school integration” (PNAS Nexus)Built React App to visualize rezoning to district leadership using Flask, Streamlit, PostgreSQL, Nginx, and Leaflet “Byte-Sized Justice: Supporting School Desegregation Policymaking with Algorithmic Redistricting” (ICCSS)

Cambridge, Massachusetts, United States

Analyzed socio-economic vs racial integration in schools, made heuristics to improve rezoning runtimes by ~10%Utilized Constraint Programming to redesign school zone boundaries in 1,000 school districts throughout the U.S. and demonstrated that school zones optimized to reduce racial segregation had lower travel times than existing zones“Redrawing attendance boundaries to promote racial and ethnic diversity in elementary schools” (ACM EAAMO)

Berkeley, California, United States

Computer Science

Mathematics

Education record