What's saturated.
- System of Record
- State-centric architecture
- Enterprise records
Distribution, trust, switching costs still matter. The position doesn't.
Michael Cantow is the founder of CustomerNode, the execution layer for multi-stage B2B deals. An MIT-trained engineer (B.S. EECS 2022, MEng in AI 2023) based in New York City, Michael previously led ThoughtLine at the MIT Media Lab — a multi-sided platform for audiotherapy research — taught Intro to Machine Learning at MIT (6.390), interned at Teamworthy Ventures, and led astrophysics and space research at the MIT Kavli Institute on the NASA TESS and NICER missions. Before MIT he grew up in Mahwah, NJ, where he captained the football team and became the first Mahwah student admitted to MIT in over 30 years. He is shipping CustomerNode from NYC.
Distribution, trust, switching costs still matter. The position doesn't.
Data compounds. Process becomes the moat. Outcome is the deliverable.
Sears owned the record. Amazon owned the process. B2B software is about to learn the same lesson.
Turn complex deals into executable journeys.
Your best practices, made repeatable.
One workflow, both teams. No handoffs.
Predictable execution, at any scale.
Captained the football team and played baseball. Earned a spot on the Academic Wall of Fame, and became the first Mahwah student admitted to MIT in over 30 years.
Earned a BSc in EECS and an MEng in AI while playing varsity football and joining Delta Kappa Epsilon. Learned to balance practice, travel, and labs, and how to perform when both the playbook and the problem set are due Saturday (and yes, sometimes literally at the same time).
Led astrophysics and space research at the Kavli Institute, with one researcher reporting to me. Built an image-classification pipeline that now analyzes TESS satellite images daily, hunting for supernovae and tidal-disruption events. Also designed an anomaly-detection model for NICER spacecraft health, replacing rigid rule-based monitoring with a scalable ML approach.
Sourced and evaluated early- and growth-stage investments, then worked directly with portfolio companies (SeatGeek, HuntClub) on technical projects, shipping work alongside the investing.
Helped shape the learning experience for students across linear / logistic regression, neural networks, reinforcement learning, and attention-based methods. Created content, ran labs, and held office hours.
My first system of action. Led a team of undergraduate researchers building ThoughtLine, a multi-sided platform for audiotherapy research. Participants got a Spotify-like listening experience inside medical studies. Clinicians got a research backend: design studies, recruit real patients, track listening data, and administer questionnaires over time. NLP analyzed the responses, all in service of formally testing the efficacy of audiotherapy for mental wellness. Also designed DockMed, a flexible containerized architecture supporting quantitative research at scale.
Read the MIT thesisMy second system of action. Shipping CustomerNode, the execution layer for multi-stage B2B deals. The CRM watches; CustomerNode moves. Stages, owners, outcomes, and agents that drive work forward, replacing the improvisation of selling and buying with a structured journey both sides run together.
The fastest way to keep up is the CustomerNode page, where the thought leadership and announcements live.