At Afresh, I was a **Founding Engineer **(now ~200 employees, ~$15M ARR).

I was a key architect building our Machine Learning, Data, Backend, Product, SRE/Infrastructure, and DevOps systems and platforms from scratch. 80+ engineers now build upon the foundations I created and the systems I designed.

Outside of the code, I shaped the company’s culture, processes, and priorities that continue to influence the company to this day.

Machine Learning

• Wrote ML pipelines for demand forecasting, inventory estimation, and ordering policy that comprise the recommendation engine (Python, Pandas, XGBoost, TensorFlow).
• Designed modularized ML architecture that underpins our entire ML system.
• Integrated ML infrastructure for serverless GPU computing, automated profiling, and performance monitoring.

Data

• Built and managed customer data integrations, ETLs, data ingestion, and automated pipelines (Python, Pandas, Airflow) for all enterprise partners.
• Designed our data models and schemas for our relational DB and data warehouse (Postgres, Snowflake).
• Built DB APIs and data platforms that serve all data operations (Python, TypeScript, Sequelize, SQLAlchemy).

Backend

• Architected our codebases (ML, Data, and Product backends), development environments (TypeScript, Python), and testing environments (integrated and unit).
• Built backend server (Node.js, TypeScript) and designed API endpoints (GraphQL) that power clients for all product features.
• Designed and built context management, error handling, and logging for both our Python and TypeScript repositories.
• Implemented type systems, extensive linting and auto-formatters, and dependency management for both our Python and TypeScript repositories.

Product

• Designed and implemented data models, data loaders, and APIs powering all client functionality, including orders and inventory.
• Designed and built the fundamental key features end-to-end, from ETL to ML to DB to APIs to GraphQL.
• Implemented fundamental backend functionality such as authentication, user and state management, and auto-scaling (Terraform).