Samay Deepak Ashar

Hey, thanks for stopping by!

I’m a software engineer drawn to the craft of systems that scale. Distributed backends, high-throughput services, cloud infrastructure that stays calm under pressure: the quiet architecture behind software that just works.

Somewhere in that work, intelligence shows up too. Models, automation, data in motion. Not as a side quest, but as part of how modern systems think, adapt, and keep moving.

If you’re into building things that last, shipping with care, or talking shop about systems, you’re in the right place. Pull up a chair.

Samay Deepak Ashar — portrait

Work

  • Software Engineer, Agentic AI

    Turing, Mercor, Handshake AI, Snorkel

    Oct 2025 – Present·Remote (USA)

    • Designed robust validation for large-scale agentic training datasets, keeping accuracy and structure steady across research programs at scale.
    • Structured 435+ JSON, Python, and Docker tasks under shared schemas so multi-format data stayed clean, consistent, and production-ready.
    • Raised dataset accuracy by 45% and cut validation cycles by 35% while keeping pipelines aligned with evolving model research objectives.
  • Data Scientist

    Launchpad

    Apr 2025 – Jun 2025·Remote (USA)

    • Built Gemini-powered automation to parse financial documents and pull live data across 15 APIs for a busy cross-functional delivery team.
    • Cut processing time by 35% and manual extraction by 55%, replacing brittle weekly workflows that once consumed more than eighty team hours.
    • Lifted data precision from about 75% to 92% through Dockerized deployment and clearer, more reliable end-to-end collection pipelines.
  • Software Development Engineer (AI)

    SchedMad Private Limited

    Jul 2024 – Nov 2024·Vadodara, India

    • Shipped a Laravel WhatsApp chatbot with containerized deployment to move customer support beyond slow, overloaded manual shift queues.
    • Tuned twelve key database queries and connected the WhatsApp Business API for steadier throughput as product traffic continued to grow.
    • Scaled support to 450+ interactions per shift with 96% delivery and brought average response latency down to roughly four seconds flat.

Startup

I believe that a startup is born from a really good project that you’ve developed over the years. It can be a college project, your research or an exciting opportunity that you have seen coming. When this project is loved by customers, it becomes a startup.

Introducing mine: Prodigy

Prodigy is an AI workforce infrastructure that provides on-demand employees based on a team’s workload and requirements. We have built our own agentic workspace and scrum environment where human and AI employees work together as part of the same team.

  • Product in Beta: reach out on prodigy.sam10@gmail.com for testing.
  • Received $175k in infra from Y Combinator and Startup Foundry.
Prodigy logo

Skills

  • Languages

    Java · Python · C++ · JavaScript · SQL · Rust

  • Backend

    Spring Boot · REST · Microservices · gRPC · RabbitMQ · Kafka · Redis

  • Frontend

    React · Node.js · Express · Tailwind

  • Cloud & DevOps

    AWS · Docker · Kubernetes · Jenkins

  • Databases

    MySQL · PostgreSQL · MongoDB · DynamoDB · Elasticsearch · Cassandra

AweSam — Sharing notes on technical and non-technical concepts I'm learning along the way.