Hi, I'm Sid 👋

Engineering lead at Hypotenuse AI, building LLM infrastructure for ecommerce product data at scale.

Work

Apr 2026 - Present

Lead Software Engineer · Hypotenuse AI

Own the architecture and direction for Hypotenuse, including the product database, enrichment, and generation systems that handle millions of SKUs every day. Most of my time goes into deciding what to build next and keeping AI and compute costs trending down as we grow.
Building the feedback loops and retraining pipelines that turn every correction across Hypotenuse, both explicit edits and implicit signals, into training data for better models over time, along with the internal AI tooling our engineers rely on day to day.

Jan 2025 - Apr 2026

Senior Software Engineer · Hypotenuse AI

Designed and shipped Hypotenuse's end-to-end product data platform: the multi-tenant database, a web crawler that uses LLMs to enrich new products as they arrive, and the bulk processor that handles 100k+ products in a single run. Today it serves sub-second queries across 2M+ products.
Built our in-browser image editor from scratch. A canvas engine with undo/redo, the usual editing tools, and AI image generation built in. Also shipped custom ProseMirror and TipTap nodes for the rich-text side of the editor.
Stood up proper observability across the product. Traces, metrics, and logs through OpenTelemetry, flowing into AWS Grafana and Prometheus, with alerts that page the right person.

Jan 2023 - Jan 2025

Software Engineer · Hypotenuse AI

Built Hypotenuse's core AI capabilities from a blank slate: RAG pipelines, LLM streaming, the vector store, rate-limiting for generation and third-party APIs, and a crawler that keeps an eye on customer-configured sites to surface fresh content ideas.
Shipped research-grounded blog workflows with inline citations, and led the platform-wide move from a credit-based pricing model to a word-based one, which touched almost every generation flow we have.
Provisioned our multi-region AWS infrastructure with Terraform. Cut image generation costs by 89% and server memory by 99.5% through profiling and targeted rewrites.

See the full work history

Selected work

All projects

project MatterIxA tiny deep learning framework, built from scratch to teach the fundamentals. pythondeep-learningautodiff 2020 → 2022 project Learnify AITurn your study notes into an interactive knowledge graph. javascriptknowledge-graphnlpedtech 2020 → 2023 project DetoxA drop-in REST API for multilingual profanity filtering. javamlapihackathon 2021 → 2023 project Newscast APIA simple REST API to fetch news articles for any query. pythonrest-apidatacrawling 2021 → 2022 project ShillingsA mock payments backend for testing client integrations end-to-end. gobackendapisystem-design 2022 project SciML for Knowledge DiscoveryA comparative study of symbolic and sparse regression for scientific ML. researchscientific-mlsymbolic-regression 2022 · ICARCV

Recent writing

All posts

post Startup Post-Mortem: Lessons from My Startup OdysseyLooking back on a startup that didn't make it — what I learned about engineering, business, design, and the small ways things tend to go wrong. startuplessonsdatlite Jan 29, 2023

Selected publications

Google Scholar

ICARCV 2022 Dec 2022

Read

A Comparative Study on Machine Learning Algorithms for Knowledge Discovery

A survey and head-to-head benchmark of symbolic regression methods — Genetic Programming, Deep Symbolic Regression, AI-Feynman, and NeSymRes — on the Feynman-03 and Nguyen-12 datasets. We highlight the strengths, failure modes, and open problems for each family of methods.

S. S. Suseela, Y. Feng, K. Mao

Nanyang Technological University 2022

Read

Scientific Machine Learning for Knowledge Discovery

Master's thesis at Nanyang Technological University. Investigates how machine learning methods can recover interpretable, closed-form models from data — bridging the gap between black-box prediction and scientific insight.

S. S. Suseela