Session Context & Core Thesis
Speaking at Tathva '26—one of South India's flagship technical festivals—Ritwik Joshi challenged the prevailing anxiety around automated code generation. Rather than treating coding agents as a threat to junior engineers, he framed them as a structural filter that exposes shallow syntax memorization.
Using real robotics and enterprise examples, the session contrasted brute-force code output with systems judgment. Engineers still own latency budgets, memory boundaries, API contracts, failure recovery, and physical actuation where AI errors carry real cost.
Students also walked through practical toolchains combining deterministic verification, Model Context Protocol (MCP) integrations, and embodied Physical AI engineering. The Q&A closed with concrete advice on building verifiable proof-of-work portfolios across Indian deeptech.