Current Role
Software Engineer 2
Guidewire Software
Bangalore, Karnataka
Jan 2024 - Present
Current focus area
Applied AI / Model Evaluation and Right-Sizing
- Evaluate smaller, specialized language models against representative workloads for task quality, cost, and latency.
- Deploy and benchmark open models including Gemma, Mistral, and Llama through internally repeatable evaluation pipelines.
- Build LLM-as-a-Judge and observability workflows, then fine-tune candidate models where task quality needs improvement.
- Shipped an internal agentic workflow that coordinates language models and custom tools for multi-step reporting.
Prior focus area within this role
Sentinel Product Engineering and Secure NL-to-SQL
- Built Go microservices for a product that analyzes software artifacts and applies security policies.
- Delivered secure natural-language-to-SQL capabilities with controls designed for enterprise data access.
- Shipped full-stack features with Go, TypeScript, PostgreSQL, AWS, and Kubernetes.
- Tracked remediation work and promoted secure engineering practices as the team's Security Champion.
Technical Expertise
Agentic System DesignLLM Orchestration (LangChain)Tool IntegrationMulti-step ReasoningPrompt EngineeringOpenAI API
PythonGo (Golang)TypeScriptSQL
AWS (S3, RDS, ECR, DynamoDB)KubernetesDockerCI/CDTerraform
Computer Vision (DeepLabV3+)Data Quality AnalysisPostgreSQL
Technologies in Production
Python, Go, TypeScript, PostgreSQL, AWS, Kubernetes, Docker, Terraform, LangChain, LiteLLM, Langfuse, DeepEval, KServe, and vLLM.
Recognition and Leadership
Google's Agentic AI Day Finalist
Selected as a finalist for Google's Agentic AI Day after building an agentic workflow under event time constraints.
Technical Speaker & Trainer
Delivered talks on Prompt Engineering and organized hackathons covering LLM optimization and AI application development.
Education Impact
Delivered training in deep learning concepts using MATLAB Deep Designer.