Rohit Punjabi

Software Development Engineer with an M.S. in Computer Science and production experience shipping C++, Python and TypeScript software across 7 live releases in a fast-paced, highly iterative Agile environment. Designed a player telemetry system on a MongoDB cloud backend from scratch and built the self-serve analytics dashboard on top of it, cutting bug investigation time by 70%.

Background in machine learning (TensorFlow, deep learning) and proficient with AI development tools such as Claude.

Software Engineer at Shrapnel (2024-02 – 2024-09)

• Starting from a broadly defined need for gameplay data ahead of early access, designed and built a telemetry system from scratch in C++ (Unreal Engine 5.3) on a MongoDB cloud backend, capturing kills, weapon usage, movement and power-up events and powering product analytics across 7 live releases.
• To shorten bug investigation for the studio, led development of an internal analytics dashboard in TypeScript and React that was adopted by product, design and QA teams and cut bug detection and analysis time by 70%.
• To remove engineering as the bottleneck for data questions, transformed raw event logs into aggregated MongoDB views queried by the dashboard, giving non-technical stakeholders self-serve access to live player data.
• Tasked with improving player engagement, designed and implemented a social system in C++ and Blueprints supporting unlimited friend requests and team play, and shipped it to live players.
• Delivered features in 7 early access releases in an Agile environment by collaborating with 5+ cross-disciplinary teams, including gameplay, backend, design and QA, across C++, TypeScript and REST API codebases.

Tools Software Engineer at Shrapnel (2023-08 – 2024-02)

• When a critical defect was destabilizing the studio's internal tooling, diagnosed the root cause and shipped a fix that reduced application crashes by 30%.
• Content review ran through a fragmented multi-step process, so built a web-based moderation tool in TypeScript and React that reduced it to a 2-step workflow for senior reviewers.
• To eliminate repetitive manual image processing, designed and implemented a Python script integrated with AWS S3 that automated the task and reduced manual workload by 30%.
• Owned client-side tool development in a 3-person team (C++, C#, Unity3D), gathering feedback from designers, artists and power users and shipping features that improved productivity by 80% and prevented data loss.

Master of Science in Computer Science – DigiPen Institute of Technology (2021 – 2023)

Bachelor of Engineering in Information Technology – Thadomal Shahani Engineering College (2017 – 2021)