Afresh is the first Fresh-first supply chain software company for grocery. Using cutting-edge technology (buzzwords like AI, ML, deep reinforcement learning, etc.), we build tools that help grocers and other supply chain constituents reduce food waste and maximize profit by forecasting demand and optimizing decisions.
We are the rare start-up that actually, credibly combines (1) Cutting-edge technology (Proof: we've been published in ICML); (2) Tangible, positive social impact (Proof: in live deployments, our system decreased food waste by >50%); and (3) High-growth potential (Proof: We've gone from Seed to Series A and signed grocery chains that represent 500+ stores and 10B revenue — all in less than 1 year!)
We have awesome investors! Our Seed stage investor, Steve Anderson at Baseline, was an early investor at Instagram, Heroku, SoFi, and Stitch Fix, and has been consistently been at the very top of the Forbes Midas List.
Food waste is a horrible, environmental blight: about 30-40% of food produced worldwide is thrown away, causing nearly a trillion dollars of economic losses, trillions of gallons of wasted water, and billions of tons of greenhouse gas emissions. In the US, about 40% of all food waste occurs at the retail level and downstream, largely driven by insufficient technology and manual processes. We are actively and tangibly helping solve this problem.
Our technology doesn’t just reduce food-waste — it makes the fresh food supply chain more efficient making fresh food fresher, tastier, cheaper, and more widely available!
You'd be joining an awesome team at the inflection point of serious growth. We have awesome people from top programs—Stanford, Cal, CMU—and our CTO Volodymyr's PhD thesis from the CS/AI program at Stanford was awarded the top thesis in his class. Yet, all of us are really friendly, kind, and helpful.
We are a flat engineering organization consists of people working on machine learning, data science/engineering, backend/infrastructure, full-stack web/mobile development, and design. We operate a lightweight Scrum process with two-week sprints and daily standups. We develop our code on Github following the Gitflow model and continuous integration and development powered by Jenkins.
We're building a decision-making system powered by reinforcement learning that optimizes tens of thousands of daily ordering decisions in hundreds to thousands of supermarkets across the US. This requires infrastructure that scales state-of-the-art model-based planning algorithms to datasets of tens of millions of data points that are ingested daily. The resulting decisions are communicated to tech-averse end-users via a mobile app. It's an end-to-end problem that has never been solved in the world of Fresh food and grocery!
Use state-of-the-art model-based reinforcement learning algorithms to automate tens of thousands of daily ordering decisions in hundreds to thousands of supermarkets across the US. Use supervised learning to create a world model, and apply planning algorithms to determine optimal decisions in that model.
Scale state-of-the-art model-based planning algorithms to datasets of tens of millions of data points that are ingested daily. Develop data pipelines powered by Apache Spark that process and clean data feeds coming from the customers and place them in our data lake.
Take the output from our reinforcement learning model and feed it into a mobile frontend that will be used every day by hundreds of store operators to plan their decisions. These generally tech-averse end-users require a UX that will fit their needs and will make them maximally effective and improve the adherence to our recommendations.
Our four cultural values are: Honesty, Proactivity, Humility and Kindness.
Unmonitored vacation policy.
Interested in this company?
Skip straight to final-round interviews by applying through Triplebyte.