Sr. Solution Architect - HCL Tech

I design agentic AI systems for IT infrastructure automation. In practice: agent workflows that ingest signals from observability and ITSM platforms, reason over them with an LLM, and hand execution to deterministic automation — BigFix, Ansible, Terraform, Jenkins, scripted runbooks — rather than generating commands at inference time. Human approval gates sit in front of anything with production blast radius. The design decision I defend most often is that split. Using the model for triage, summarization, and routing while keeping remediation deterministic trades away flexibility and buys a bounded failure mode: a bad classification costs a misroute, a bad generated command costs an outage. Retrieval over runbooks and historical incidents grounds the reasoning layer, and cases without supporting evidence abstain and escalate rather than guess. Prompts are versioned and regression-tested against a held-out incident set before anything ships. Alongside that, I architect BigFix-based solutions across the HCL software portfolio — endpoint lifecycle and patch management, Runbook AI automation, cloud lifecycle management, ITSM integration, job scheduling, and reporting and business observability. I position the suite in customer engagements and then extend it, building agentic automation on top of the BigFix APIs so the agent layer has a real execution surface rather than a demo one. The foundation under all of it is a consolidated signal plane: Splunk, Dynatrace, Datadog, New Relic, OpsRamp, Grafana, and ServiceNow, stitched together with custom event-correlation pipelines in place of fragmented legacy tooling. Alert noise reaching on-call dropped by [40%], and recurring manual toil. I also lead solution architecture and pre-sales for these engagements — discovery, technical proposal, demo, delivery handoff — run release and upgrade programs for live customer deployments, and review designs before they reach the customer.