Personal details
Challa R.
Technical skills
About
I'm an engineering leader with over 20 years of experience, including 10+ at Microsoft. My work spans telecom protocols at Nokia, regulated clinical systems at Roche Diagnostics, PLM at Siemens, and now enterprise AI. Today I set the platform and AI direction for Microsoft Sustainability Manager and architect its agent harness, the governed path that connects conversation to model to enterprise data. Throughout my career I've focused on tying tactical engineering decisions to business outcomes. One example is a tracing approach that cut support SLAs by 60% and became the org-wide standard. Looking ahead, I want to grow into a principal or architect-level role shaping how organizations build and govern AI agents at scale. I'd like to define platform strategy, set standards that multiple teams adopt, and make sure responsible AI is built into the architecture rather than added afterward.
Work Experience
Platform fundamentals. Owns the layer the team builds on — PCF / Fluent UI through C# APIs and Dataverse; Scope 1–3 ingest, factor mappings, calculation profiles.
Harness engineering pipeline, end to end. Semantic Kernel holds memory state and chat history; Azure API Management applies guardrails and rate limits on every call; frontier LLMs run on Azure AI Foundry serverless endpoints; MCP servers on Azure Functions / Web Apps (SSE / HTTP) expose enterprise SQL and Microsoft Graph as tools.
D365 F&O. Account, payment and ledger flows (X++): entitlement of a transaction to post, reconciliation across ledgers.
Tracing that became the standard. Correlation IDs across a front-end / back-end boundary that used to swallow tickets. Behind the 60% support-SLA cut on page 1.
F&O banking and analytics. Payment and ledger entitlement in X++. Metrics layer on Azure Data Lake — live dashboards instead of a weekly export.
Platform and product. Modular .NET APIs, monorepos, CI/CD, TypeScript surfaces. Office 365 on React/Redux and ASP.NET OData. Dynamics 365 Community Portals (Ideas). Early Bot Framework work.
Localization as a pipeline. Partnered with Contentful so new locales shipped through content and CI, not a per-market engineering project. Owning US-team deliverables from India since 2017.