Bhargava's Profound AI Rep
Bhargava turns customer signal into reliable AI infrastructure, from call audits to observability dashboards to scalable prompt architecture.
Bhargava T
Edges
Building AI Reliability Infrastructure
Bhargava builds the observability and eval layers that make AI products reliable enough to operate in high-stakes environments. His focus is on the unglamorous infrastructure: unified monitoring, threshold alerting, and the feedback loops that catch failures before they become downtime. He did this at SquadStack, where he unified fragmented telemetry across seven voice AI systems into a single dashboard and drove campaign reliability above 95%. He did it again at Keka, where he built the eval and A/B testing framework that took a GenAI chatbot from demo to production grade. The mechanism is consistent: instrument the system first, make failures visible, then iterate on the product layer with confidence.
Spotted in 2 Stories
Scaling AI Agents Through Iteration
Bhargava scales AI agent deployments by running structured iteration loops grounded in real customer signal. He audits calls, identifies the gaps the model cannot see, and feeds that back into the product before pushing for volume. At SquadStack with Kotak Bank, he ran phased iteration across zero-to-one, one-to-ten, and ten-to-hundred stages, auditing calls at each phase to find the objections the agent was missing. At Keka, he applied the same discipline to a GenAI chatbot, using A/B evals and weekly stakeholder cadences to move from demo to production. The result in both cases was a product that could handle real-world load because the failure modes had been found and fixed before scale was attempted.
Spotted in 2 Stories
Driving Stakeholder Adoption of AI Products
Bhargava earns stakeholder trust for AI products by showing metrics before asking for commitment. He demonstrates what the product does well, names what it does not, and builds conviction through visibility and cadence. At Keka, he aligned the customer support leader on the GenAI chatbot by running a transparent demo, then maintained weekly and bi-weekly cadences until the team committed resources. At SquadStack, he kept the engineering team aligned on the observability build by surfacing the builders' daily complaints until the severity of the problem made the investment obvious. In both cases, the adoption came from making the problem and the solution visible at the same time.
Spotted in 2 Stories