Purshotam Shah

Harness engineering is the production substrate that sits between an LLM agent and your live cluster: structured audit on every tool call, per-tool quotas, shadow mode, dry-run, human-in-the-loop approval for risky writes, trace propagation, retries, and a closed feedback loop from human PR review back into the agent's retrieval. It is what separates “we demoed the agent once” from “we left it on for six months and it kept the on-call asleep.” As streaming operations adopt autonomous agents, the harness — not the model — is the load-bearing engineering. We built and run an AI debugging agent for Apache Flink. It ships as a JobManager sidecar in every Flink application-mode deployment: loopback to Flink REST, job-scoped Kubernetes permissions through a namespaced ServiceAccount, autoscaler state read from PostgreSQL, and inherited service identity for outbound calls. Per-deployment blast radius. Per-deployment audit trail. No central agent service, no fleet-wide credentials. The diagnostic agent is a LangChain ReAct loop over eighteen Flink-specific tools: backpressure, checkpoint health, Kubernetes events, autoscaler state, thread dumps, and log search. It picks two to five tools…

How do you keep your developers effective when your internal Flink practices diverge from the open-source community? At Yahoo, we faced this challenge by building FlinkAI, a smart knowledge system that bridges the gap between our internal expertise and the global Apache Flink community. In this session, we'll show you how we use Apache Flink itself to power a real-time streaming pipeline that ingests, processes, and understands Flink knowledge. FlinkAI consumes everything from our internal deployment guides (EKS, mTLS, Okta) to external community firehoses like mailing lists, Jiras, and commits. Using OpenAI, this data is transformed into semantic embeddings and stored in a vector database for lightning-fast natural language search. The best part? It’s integrated directly into the Flink Web UI. FlinkAI automatically analyzes exceptions as they happen and suggests solutions in an embedded chat, turning the UI into an active troubleshooting assistant. Come to this session to learn: A novel architecture for a streaming-first, LLM-powered knowledge system. How to leverage Flink to build powerful internal tooling for developers. Practical lessons on integrating LLMs and vector…

At Yahoo, we built a secure, scalable, and cost-efficient batch processing platform using Amazon MWAA to orchestrate Apache Flink jobs on EKS, managed by the Flink Kubernetes Operator.