Open to opportunities

KRISH SHAH

@krishshah2

Message

I build production Python, AWS, and GenAI systems for reliable real-world workflows.

India

Message

What I'm looking for

I'm looking to build reliable production software and GenAI products where I can own systems end to end, from requirements and architecture through testing, deployment, observability, and production support.

At Sunflower Labs, I build a multi-tenant SaaS communication platform handling 10,000+ patient communications monthly across SMS, email, and voice. I ship schema-validated GenAI agents, LLM evaluation and observability with Langfuse, and event-driven AWS pipelines designed to prevent duplicate patient communications.

I've also built InsightOps, reducing report query time by roughly half for 50,000+ records, and StreamPulse, a real-time Kafka pipeline sustaining 1,000+ transaction events per second. I work end to end across Python, FastAPI, PostgreSQL, AWS, React/Next.js, CI/CD, testing, and production support.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer

Current

Sunflower Labs

Sep 2025 - Present (11 months)

Design and maintain a multi-tenant SaaS communication platform handling 10,000+ patient communications per month. Shipped GenAI voice and messaging agents with structured output and function-calling contracts, and built the LLM evaluation and observability layer.

Python fastAPI PostgreSQL AWS LangGraph LlamaIndex RAG Function Calling Structured Outputs Prompt Engineering LLM Evaluation Langfuse React Next.js GitHub Actions Pytest Event Driven Architecture RBAC

Software Engineer Trainee

Sunflower Labs

Jan 2025 - Aug 2025 (7 months)

Built InsightOps, a cloud-native analytics service on AWS serving 50,000+ records, and reduced report query time by roughly half. Developed a scikit-learn risk-scoring model for churn prediction and built automated evaluation pipelines.

Python fastAPI AWS scikit learn Ragas LLM As A Judge Docker CI CD Agile Scrum

Software Engineer

Freelance

May 2024 - Dec 2024 (7 months)

Built StreamPulse, a real-time event ingestion pipeline sustaining 1,000+ transaction events per second with anomaly detection. Containerized on AWS EC2 with integration tests covering ingestion and alerting paths.

Kafka PostgreSQL Redis AWS Docker Python