Senior Software Developer

• Driving Harness Engineering initiatives through AI-assisted development workflows, evaluation mechanisms, and guardrails to improve code generation accuracy and developer productivity.
• Designed a hierarchical multi-agent orchestration framework using Google ADK, introducing supervisor agents for tool governance, interview flow control, and candidate evaluation.
• Led the evolution of a real-time AI interview platform from a custom WebSocket architecture to a WebRTC stack powered by LiveKit and Google ADK.
• Built an on-the-fly AI evaluation framework that continuously monitored agent behavior, tool usage, and interview progression in production.
• Architected and operated GPU-backed AI infrastructure on GKE Autopilot for speech and inference workloads.

Software Developer

• Spearheaded migration from Daily.co to a self-managed LiveKit deployment supporting 350+ concurrent sessions while reducing infrastructure costs by ~75%.
• Achieved ~30x speech-processing cost efficiency by replacing external speech APIs with self-hosted inference on GPU-backed Kubernetes workloads.
• Designed autoscaling infrastructure using Kubernetes HPA and KEDA to dynamically scale communication, speech, and AI workloads during burst traffic.
• Implemented prompt caching and inference optimization across hosted LLM providers to reduce latency and operating expenses.
• Conducted load testing and performance tuning of distributed AI systems, resolving scaling bottlenecks in high-concurrency workloads.

Full Stack Developer Intern

• Built React and Django applications for an AI-driven interview scheduling platform and recruiter workflow automation.
• Developed LLM-powered resume screening workflows and implemented automated backend testing using Pytest.