- Co-architected a React-based unified application providing a real-time interface for cross-team users to monitor large-scale data processing pipelines.
- Led the architecture of an E2E automated development workflow ingesting requirements from Jira, driving development, and integrating automated validation — reducing manual handoffs across the SDLC.
- Co-designed an orchestration layer with Google ADK + AWS Bedrock (Claude Sonnet 4.5) to synthesize responses from multiple heterogeneous data sources.
- Designed a distributed ingestion pipeline orchestrating multi-source data loads and embedding generation using managed AI services.
Shreesh Srivastava
I live in servers, and build cool stuff.
Software Engineer II at JPMorgan Chase & Co. I architect AI/ML data platforms, distributed systems, and GenAI orchestration — and I >_ keep shipping.
Hey, I'm Shreesh
I'm a Software Engineer II at JPMorgan Chase & Co., where I build AI/ML data platforms and distributed systems used across the bank. My work spans GenAI orchestration with Google ADK and AWS Bedrock, real-time observability apps in React, and moving large-scale batch workloads to the cloud.
I care about the whole arc — from system design and the architecture decisions that keep a platform reliable, to the interfaces people actually use. When I'm not deep in distributed systems, you'll find me on hiking trails, reading comics, or with a cup of coffee.

Where I've been building
- Architected a distributed reconciliation framework in Python validating data across files and databases; integrated an AWS SageMaker ML model to detect discrepancies.
- Migrated large-scale batch workloads from on-premises infrastructure to the cloud, cutting infrastructure costs while preserving reliability.
- Built a React self-service reporting tool across multiple data sources, removing significant manual production support work.
- Migrated a high-frequency data feed pipeline to the cloud (managed serverless services), improving scalability and cutting processing runtime.
- Automated Python diagnostics for real-time data feed validation, significantly reducing troubleshooting time.
- Architected a position aggregation platform (Spring Boot + Angular 14), owning the full lifecycle from system design to production.
- Built a business intelligence interface for operational monitoring, offloading much of the routine SRE workload.
- Redesigned the frontend with micro-frontends (Webpack Module Federation + React), improving load performance and modular scalability.
- Automated 5+ high-frequency operational workflows in Java, reducing manual ops effort.
My toolbox
Things I've built
Computer-vision pipeline that detects road lane lines with 95% accuracy across varying lighting conditions — research published in a Springer conference paper.
Source ↗Hybrid recommender blending collaborative and content-based filtering with caching, achieving 89% accuracy on IMDB data.
Source ↗Let's build something cool together
I'm currently open to interesting opportunities in AI/ML platform engineering, distributed systems, and backend engineering. My inbox is always open.