§ 00 — introduction
rev. 2026.04 — v1.0

aneesh pothuru

positioning

software engineer at waymo, building data exploration and eval systems for ml research. co-founded bench ai.

based
san jose, ca
contact
pothuruaneesh23@gmail.com
chapter 01
experience
§01

index of work.

four roles, sorted newest first. tap any row to expand.

0 entries
chapter 02
case studies
§02

selected work.

a case study from waymo, then bench ai (the company i co-founded), then a school capstone. the numbers are good-faith estimates.

0 entries
featured · case study 2025 — present

one hundred scattered datasets,
one place to look.

How long does it take a researcher to go from a hypothesis to a training run? At Waymo that gap is measured in hours, and the systems intelligence platform exists to shrink it. I work on the exploration and evaluation layer: the tooling researchers use to find, slice, and validate every dataset the company has collected, and to trace where each one came from.

0%
adoption on shared datasets
0%
dev-time reduced by agent scans
0+
insight tools consolidated
0
unified research index

what i shipped

  • 01
    Centralized every Waymo dataset into one shared index and built the exploration tooling around it.
  • 02
    Added semantic search, validations, visualizations, and dataset lineage to the insights platform.
  • 03
    Built the ETL pipelines that turn raw data into training sets and validation reports.
  • 04
    Built AI agents that scan services for vulnerabilities and operational gaps, cutting discovery time by about 90%.
  • 05
    Hooked up MCP and agent skills so researchers can request artifacts in plain language.
  • 06
    Gave auto-labeling teams real-time batch and stream views of label-collection quality.
company · bench ai

automating cloud for ml.

co-founded an iaas + saas company that gave ml teams a one-click cloud. raised institutional capital, shipped the platform end-to-end.

Built the platform from the infrastructure up: CI/CD, Nginx, Docker, AWS. Then tuned the backend until Lambda durations dropped enough to cut compute costs by 30%. The hardest part at seed stage was deciding what not to build.

what shipped

  • 01
    CI/CD pipelines, an Nginx server, Docker images, and our AWS setup, all from scratch.
  • 02
    A tuned Spring Boot backend; the shorter Lambda runs cut costs by 30%.
  • 03
    A React UI with code-splitting and SSR for faster loads and better SEO.
  • 04
    A Django layer that connects the public and internal APIs.
role
co-founder, ceo & engineer
tenure
may 2022 — nov 2023
stack
react spring django aws docker nginx postgres
outcome
institutional capital · platform shipped
project · spacepark

a 15-person parking garage.

a portable hardware + mobile app stack that detects traffic flow in parking garages — object size, velocity, direction — streamed over wifi to a web server.

Led 15 engineers as project director — mechanical, electrical, and software people who don't usually build together. Most of my time went to translating between CAD, firmware, APIs, and the app.

what shipped

  • 01
    A mobile app that reads current occupancy and expected-frequency curves for any garage.
  • 02
    A portable sensor device that detects object size, velocity, and direction; streams over wifi.
  • 03
    A full execution strategy — WBS, Gantt, technical diagrams — that kept the team shipping weekly.
role
project director / engineer
team
15 engineers
stack
react native firebase arduino cad solidworks figma
year
final capstone, sjsu
chapter 03
stack
§03

skills.

grouped by what they do, not how popular they are. the tools below show up across the work above.

three groups
languages

java, javascript, typescript, python, c++.

infrastructure

aws, gcp, grpc, spanner, apache beam, docker, nginx, spring, react.js, react native, django, postgres, dynamodb, lambda.

machine learning

ml platform, distributed systems, mcp, rag, post-training, agent skills, etl, data lineage.

chapter 04
contact
§04

always up for a good problem.

currently at waymo and not actively looking, but always reading email. for collaborations, advice, or interesting problems —

six lines
emailpothuruaneesh23@gmail.com
linkedin/in/aneesh-pothuru
github/Aneesh-Pothuru
phone(209) 637-6128
basedsan jose, ca
résuméresume.pdf ↓