Salviya Balami.

What I worked on

• Terraform drift detection. Automated comparisons between infrastructure-as-code and deployed Azure resources across 40+ cloud environments, while filtering execution failures and non-actionable differences.
• Actionable remediation. Connected real drift findings to GitHub-based workflows that turn discrepancies into issues and remediation pull requests, reducing manual investigation for cloud engineers.
• AI Innovation Tiger Team. Worked on a 30+ member cross-functional team evaluating multi-agent architectures for retrieving and analyzing engineering designs, documenting tradeoffs, and demonstrating MVPs to stakeholders.

What I worked on

• Real-time teleoperation. Built Python and ROS tooling that translated tablet control into joint commands for a scaled SPOT robotic arm during pick-and-place tasks.
• Hardware integration. Worked with Dynamixel servos, embedded hardware, CAD, and 3D-printed components to assemble and improve a low-cost manipulation platform.
• Imitation-learning data. Collected hundreds of teleoperation demonstrations containing visual observations and joint actions for downstream learning experiments, then presented the work in a research poster.

What I explored

• Trading and markets. Worked through trading fundamentals, market-design ideas, and exercises built around decision-making under uncertainty.
• Functional programming. Used OCaml in technical sessions and gained exposure to the programming style and tools used in Jane Street’s environment.
• Quantitative problem solving. Practiced estimation, structured reasoning, and collaborative problem solving with other participants and Jane Street engineers.