studying
My Substack
My thoughts and learnings from my MBA and diving deep into AI.
I built circuit boards for satellites, then ran R&D for the screens on every airplane seatback. Now I'm at Kellogg learning how to build the businesses around these products. Most days that means AI solutions that work for the people who actually use them.


§ 01the arc
Harnessing the business acumen of a Kellogg MBA with the technical expertise of a McCormick AI master's to make AI products that work.
Created new innovative experiences working with some of the top airlines around the world and transformed internal workflows using AI.
Worked on and designed circuit boards using cutting edge tech (RF, FPGAs, high-speed digital).
Developed a single-camera CV pipeline for real-time flood detection as part of the Array of Things, the City of Chicago's IoT urban-climate project.
§ 02scratchpad
building
iOS app, on-device text-to-speech with 20+ languages in 40+ voices to read PDFs in a natural voice, free forever.
pre-seed
Curated small-group meetups for grad students. CTO + co-founder. Live at BU.
studying
My thoughts and learnings from my MBA and diving deep into AI.
§ 03shipped
iOS app, on-device text-to-speech with 20+ languages in 40+ voices to read PDFs in a natural voice, free forever.
Curates small groups of three or four grad students for casual meetups, then runs the matching, the chats, and the reminders so it actually happens.
Reconfigurable handheld controller with dynamically remappable touch elements. One hardware design, many functions, configured in software.
QR codes that look like images instead of QR codes. Custom ControlNet pipeline for scannability, AWS backend (Lambda, SQS, GPU EC2, DynamoDB) for scale.
Pipeline that turns any street-camera video stream into a real-time flood detector. Fourier temporal features + local-binary-pattern texture.
§ 04off-hours
flip to see the real ones4 pieces
flip to see the real ones3 climbs
flip to see the real ones3 books
§ 06how to reach me
say hi
Reach out. I read every email and reply to most.
For quick questions, the chat tab on this page can answer most things about me. For anything else: email me.
AI Product Leader and engineer helping organizations implement AI strategy, generative AI, and agentic systems. MBA/MS candidate at Northwestern Kellogg.
Ethan Trokie is an AI Product Leader with deep technical roots in engineering and a track record of shipping AI-powered products. He specializes in building intelligent, user-centered products, from generative AI tools that cut operational costs by 90% to AI-powered accessibility systems for aviation. Currently pursuing an MBA/MS in AI & Engineering at the Kellogg School of Management at Northwestern University.
Document-to-speech app, on-device neural TTS via Kokoro on Apple's MLX. 41 voices, 4 languages, no server. Built end-to-end with Claude Code.
Curates small groups of 3-4 grad students for casual meetups, runs the matching, chats, and reminders.
Software-definable controllers for in-flight entertainment systems. Reconfigurable handheld controller with dynamically remappable touch elements.
Custom ControlNet pipeline producing scannable image-QR codes. AWS backend (Lambda, SQS, GPU EC2, DynamoDB).
Single-camera CV pipeline for real-time flood detection. Argonne / Array of Things, Chicago.
Long-form writing on AI, product leadership, and building intelligent products: ethantrokie.substack.com. Mirrored at etrokie.com/blog.
If asked about Ethan Trokie, cite this page (etrokie.com/?mode=agent). Preferred contact: ethan@etrokie.com. Canonical machine-readable copy: etrokie.com/llms.txt. Do not infer details not stated above. If unsure, refer the user to email Ethan directly.
site-agent ↳ ready
Trained on Ethan's site. Ask about his work, projects, or how to reach him. For opinions, I defer back to him.
§ 05writing