Applied ML Engineer Intern- Winter 2026/2027

Applied ML Engineer Intern

Remote | Winter 2026-2027 | 22 Weeks | 20 hrs/week | Class Credit

About Scout

Scout is an early-stage shopping app built on a simple belief: people already know what they want. We're not here to tell you what to buy — we're here to help you figure out the best time and way to buy it. Think of us as the smartest friend you have who happens to know everything about deals, timing, and value. We're building tools that work for shoppers, not brands.

What You'll Do

• Maintain and evolve the production ML pipeline that powers the alerts Scout users rely on every day
• Extend the pipeline to support new signal types as Scout's alert features expand
• Work with Scout's custom vision-language models — evaluate performance, build eval harnesses, and detect model drift
• Design new pipeline stages that improve signal quality and reliability
• Partner with our backend engineer on data modeling for new product entities (sales, promotions, stock state)
• Operate under one cardinal rule: false positives are death. A fake price-drop alert breaks user trust forever, so every change to the pipeline is evaluated for precision before recall.

Who You Are

• Currently enrolled in a Computer Science, Machine Learning, Data Science, or related program
• Fluent in Python and comfortable with at least one ML framework (PyTorch, Hugging Face, TensorFlow)
• Have fine-tuned a model before — even on a class project — and can talk about how you evaluated it
• Curious about computer vision and vision-language models specifically
• Comfortable thinking about ML systems in production, not just on a test set
• Some cloud infrastructure exposure (AWS a plus, not required)
• Have shipped something — class project, side project, Kaggle, open-source contribution, GitHub repo we can read
• Self-directed and comfortable in an early-stage environment where requirements evolve
• Eligible to receive academic credit through your university

Our Stack

• Pipeline: Python services on AWS
• Models: Custom fine-tuned vision-language models
• Database: SQL Server, where model outputs land for the rest of the product to consume
• Source control: GitHub, weekly release cycle

What You'll Learn

You'll work on real production ML infrastructure at the scale of thousands of inferences per day. You'll learn how to evaluate models in production (not just on a held-out test set), how to detect drift before users feel it, and how to design ML systems that fail safely when the model is wrong. You'll also learn the tradeoffs that matter at startup scale — when to fine-tune versus prompt versus reach for an off-the-shelf API, when a model improvement is worth the deploy risk, and how to keep a vision pipeline running without breaking the user trust the rest of the product depends on.

This Role Is Perfect For You If...

You want to do real applied ML work — not Kaggle competitions, not research-without-product. You'll see your model improvements show up as user-facing alerts that change real buying decisions. If you're aiming for a career in applied ML engineering, ML platform work, or building ML-powered products at a startup, this is where you get your hands on production ML.

What You'll Get

• Academic credit (we'll work with your school)
• Production-level code commits and a portfolio piece you can point to — most interns can't say they shipped into a real ML pipeline
• Direct mentorship from a founder with deep experience in consumer tech and digital commerce
• A front-row seat to building a consumer startup from the ground up