User's Cover Picture
User's avatar

Rohan Raj

Message
Intro
Brooklyn, United States
Graduate Research Assistant at RiskEcon Lab for Decision Metrics, NYU
Studied Computer Science at New York University
Studied Mathematics and Computing at Delhi Technological University
Joined March 23, 2021

Skills

About
Masters in Computer Science at NYU Courant | Actively Seeking May 2021 Full-Time SDE/SWE and Data Science Opportunities
Experience
User avatar
RiskEcon Lab for Decision Metrics, NYU
Sep 2019 – Present
New York, NY
Graduate Research Assistant
• Introduced a novel pipeline for Automatic License Plate Recognition (ALPR) of US License Plates with 5 separate convolutional neural networks. • Slashed prediction time by 40% and increased prediction accuracy by 75% of Automatic License Plate Recognition using Convolutional Neural Networks.
User avatar
Numerati Partners LLC
Oct 2020 – Dec 2020
New York, NY
Research Associate Intern
• Led a team of 6 researchers for peer review, technical review and other related research and evaluation assignments using Python. • Conducted hacks on cloud-based detectors through black-box adversarial attacks using AWS S3 buckets.
User avatar
PricewaterhouseCoopers (PwC)
Jul 2018 – May 2019
Gurugram, India
Data Scientist
• Reduced average throughput time by 3 days, slashed rework rate by 20% and identified possible steps for increasing automation rate for P2P and O2C processes for a leading packaged foods company. • Improved efficiency by 50% and reduced chances of fraud by developing Python code to automatically watermark customer documents for leading electricity DISCOM. • Dashboard, report development & business analysis of P2P and O2C processes using process mining techniques to generate actionable insights for Fortune 500 clients. • Served as the section leader and mentor for Python training sessions for a group of 35 employees in the Data and Analytics business unit of the firm.
User avatar
Computer Vision Center, UAB
Dec 2017 – Feb 2018
Barcelona, Spain
Research Intern
• Implemented Domain Adaptation for the ELEKTRA Autonomous Vehicle project where the network is trained on the CARLA simulator and tested on the St. Quirze Circuit, Barcelona. • Increased track completion rate by 6% by reprogramming control for ELEKTRA and CARLA simulator to take continuous steer input. • Introduced new evaluation technique which penalized wrong heading and coordinates of ELEKTRA.
Education
User avatar
New York University
Sep 2019 – Present
Masters in Science, Computer Science
User avatar
Delhi Technological University
Sep 2014 – May 2018
Bachelor of Technology, Mathematics and Computing