Architecting and scaling core backend capabilities for Zia Agents, building platform-level features end to end.

• Built the backbone for agent orchestration, shared state, context management, and multi-agent workflows.
• Took ownership & led a major architectural shift from synchronous, blocking flows to a distributed event-driven architecture that eventually enabled new platform-level capabilities like multi-agents collaboration, observability, long-running agents, etc, while improving fault tolerance and independent scaling.
• Set the standard for designing structured, agent-friendly APIs to be consumed by clients, internal services, and end users.
• Improved observability and reliability through better tracing, queue-aware processing, and resilient service patterns.
• Managing the application's release & deployment cycle and successfully handled multiple migrations with zero downtime.

Delivered end-to-end MLOps features for no-code machine learning workflows.

• Shipped key features like Data Visualization, ML Pipeline Templates, and Cross-DC Sync.
• Implemented RAG Knowledge Base, LLM Serving and internal platform utilities to improve onboarding and DX.
• Worked full-stack on various features end to end
• Collaborated across product and engineering teams to launch scalable platform features.

Learning experience by building hand-on projects with increasing complexity.

• Incrementally built a full-stack cricket score tracking application `CricketDugout`
• Implemented various features like Live Scoreboard, Management Views, Custom OAuth based Authentication, etc
• Worked on minor features in Catalyst QuickML
• Leart a lot about the internal frameworks and the way of work