open to work: software, ml, and data engineering

i build things that keep working after the demo.

hi, i'm Suhrith. i work across backends, data pipelines, and ml, with an ms in computer and information science from purdue.

right now:

pr merge predictor, self-verifying next check in 05:59:59
0.923auc at training
hover the line
last verified 0m ago

sample data, shown for illustration.

0order lines through a dbt-tested pipeline
0recall tracking vehicles on kitti
0employer records quality-scored

my stack, as a graph

hover a node

poke around

about

ms cis, purdue, may 2025 btech it, manipal jaipur software engineer intern at drdo

i'm a backend-leaning engineer who works across data, ml, and full-stack. at drdo i built python models for target position prediction, and since then i've built a portfolio that runs from vision pipelines to warehouses to llm apps.

the thread through all of it is checking my own work: point-in-time features so a model can't peek at the future, dbt tests on every model, and a predictor that re-verifies its accuracy on a schedule. i'm most interested in backend engineering, distributed systems, and agentic ai.

journey

scroll down the line. click a stop to open it.

    • graduated may 2023.
    • where the software and data fundamentals came from.
    • built python predictive models that reached 90% target position accuracy.
    • cut analytical workflow time by 40% using matlab and python.
    • packaged python and c++ apps with docker for teams on different platforms.
    • turned analytical requirements into data pipelines using tdd and oop.
    • graduate degree in computer and information science.
    • after graduating, i built out the projects below on my own.
    • shipped a self-verifying merge predictor, a job-matching engine, and vision and data pipelines.
    • looking for a team that cares about reliability.

projects

filter by kind of work. open a card to see how it's built.

build, ship, verify

pick a stage to see the tools behind it.

say hi

tell me what you're building.