Lalit Yadav

I ship software people actually use:

A live AI interview platform with 350+ users.
Code merged into Apache Beam and ClickHouse, data engines used by thousands of companies.

Highlights

350+ users

An AI interviewer you can actually talk to. I built it, launched it, and run it in production.

6 merged PRs

My code ships in Google's open-source engine for processing data at massive scale.

2 merged PRs

Contributed new functions to one of the world's fastest analytics databases.

Software Engineering Intern
May 2026 – Present
Building Construction Science, MSU
Fine-tuning Qwen3-4B + XGBoost for construction bid price prediction over 1M+ historical bids: 12.4% MAPE, beating baselines by 18%.
Software Engineering Intern
Jan 2025 – May 2025
Center for Advanced Vehicular Systems
React dashboard for real-time control of autonomous vehicle fleets with sub-200ms latency over WebSockets and ROS.
Software Engineering Researcher
Dec 2023 – Sep 2024
Wireless Communication Lab
Simulated 5G networks with srsRAN and Open5GS; moved pipelines to Kafka to stream 25M IQ samples/sec at 85% lower latency.
Software Engineering Intern
Jan 2024 – May 2024
Center for Advanced Vehicular Systems
Improved a full-stack MERN survey platform and automated analysis of 100+ survey responses.