Hey, I'm Max Poff
I'm a Master's Student at MIT, passionate about Applied AI, and hoping to shape how the world adapts to these new technologies. With 4 years of experience leading manufacturing operations at P&G and Tesla, I've built a strong foundation in operational excellence. I have a successful track record managing teams under uncertainty and driving a culture of continuous improvement. I'm excited to combine this practical experience with the many new technical skills I've learned at MIT to create meaningful impact in my next role. 🚀
Project Portfolio
Retrieval-Augmented Generation
I built a full-stack web application (see repo here) that implements RAG from scratch, using a React/Next.js frontend and FastAPI python backend. I deployed the frontend on Vercel and the backend on Render.
Check it out here! -> https://rag-pipeline-one.vercel.app/
↳ ⚠️ Warning! The latency is pretty painful on Render's free tier.
Computer-Using Agent
As part of MIT Sloan's GenAI Lab, I'm using OpenAI's Computer Use Agent (CUA) API to autonomously navigate our sponsor company's website as part of a broader pipeline to asses their customers' user experience.
Modeling Airport Gate Assignments
Under supervision of Professor Alexandre Jacquillat, I'm working on scalable optimization algorithms that assign flights to gates in a way that minimizes walking distances for passengers with tight connections. No more sprinting across terminals 🙏
Multimodal Deep Learning
As a course project for 6.7960 at MIT, I designed a multimodal neural network architecture and trained it end-to-end to predict the sale price of a home based on a picture of it's exterior (curb appeal) and it's tabular features (sq ft, bedrooms, bathorooms, etc). As you might expect, curb appeal matters to buyers and I show that multimodal models outperfrom xgboost trained on tabular features only. Read my full blog post here that explores the different ways to incorporate image data into ML models.
YouTube Channel
Teaching has always been a passion of mine. I created a YouTube channel called STEM Support (with the right amount of clickbait titles) while I was a linear algebra TA at Georgia Tech. Off to the right (or below on mobile) is some of the support the channel has received over the years!
