Founding AI Research CTO

83 Sciences is a YC-backed AI-for-materials company turning academia's unpublished experimental data into new, synthesizable materials. It was founded by three co-founders who met through an MIT commercialization fellowship: CEO Ian Naccarella (ex-BCG; previously at a silicon-anode battery startup), CTO Eric (whose PhD built a tool that determines a crystal's structure from its powder X-ray diffraction pattern, the technical seed of the company), and COO Yang Kong (a former BCG colleague who led BCG's AI implementation practice).

The insight: frontier materials models are good at proposing new materials but bad at proposing ones that can actually be synthesized, and the missing ingredient is real-world experimental data, which sits unused in lab notebooks across academia. 83 Sciences pulls in that data through direct professor collaborations, already 20 of them, deliberately capped because the process is high-touch, and uses it to drive novel materials discovery, proven out first with a University of Utah collaboration that produced a new material.

The commercial wedge: selling discovered "co-product" materials (sorbents, additives, binders) to R&D directors and VPs at inorganic-chemicals and materials companies, the necessary-but-not-core-IP inputs those companies would rather buy than build, like a sorbent that removes impurities upstream to lift a customer's copper-extraction yield. The team is raising a several-million-dollar seed around (Sept 10) and moving to New York shortly after.

83 Sciences is looking for its 4th co-founder and CTO, clear first choice over a founding engineer. This person takes the CTO seat and owns the AI/ML engineering and the entire data-and-discovery engine, freeing current CTO Eric, who is very strong on the science, to move into a chief-science / application-focused role. Because a several-million seed is closing shortly and the plan is to hire engineering talent fast after it, the co-founder mandate explicitly includes recruiting and building the team, so a strong network in AI-for-materials is a real asset.

For a genuine rockstar who isn't ready for co-founder risk, there's a founding-engineer path that can grow into CTO. Concretely, the first ~6 months: Month 1, build high-throughput, standardized data-ingestion pipelines to pull in 20+ labs' unstructured experimental data. Months 2-3, demonstrate model-performance uplift: benchmark models with vs. without the experimental data to produce hard proof points. Months 4-6, run the materials-discovery pipeline end to end to surface novel, commercially relevant materials the founders take to industry, while hiring out the team below them.

It's hands-on infrastructure and modeling first, paired with the team-building and partnership judgment of a co-founder.