Poindexter Labs Secures £2M Seed Round Led by Episode 1 to Build Expert Training Data for Frontier AI

Poindexter Labs argues that this legacy model is the weak link in the AI supply chain, calling it a "broken" pipeline for frontier AI . The company focuses on producing what it calls "high-fidelity training and evaluation datasets" that incorporate proofs, chain-of-thought traces, multi-hop logic, and detailed step logs across STEM, legal, medical, finance, and engineering domains .

Syncronus: A peer-review platform instead of a labeling conveyor belt

The company's technical answer is a platform called Syncronus, which replaces the isolated annotator model with a structured, collaborative peer-review process . Instead of a single contractor completing a task in isolation, problems are created and then reviewed by a vetted network of Olympiad medalists, PhDs, and professors .

A typical task on the platform might involve authoring an original computer science problem that requires a multi-step proof. The solution is captured with full process evidence—scratch reasoning, step logs, and LaTeX diffs—which is then reviewed by a separate expert for correctness and clarity . This creates a paper trail of expert cognition that can be used directly for instruction-tuning sets and reasoning curricula, or for running small to medium-scale model fine-tunes .

Poindexter licenses the Syncronus platform to enterprises and government bodies that want to create their own curated expert datasets. It also runs an in-house data annotation service that delivers finished, peer-reviewed datasets directly to frontier AI labs .

Where the funding will go

The company plans to use the new capital primarily to accelerate development of the Syncronus platform and to scale its contributor network . As demand grows from AI labs that need high-quality reasoning data for both training and evaluation, the company is betting that its model—combining platform technology with an elite human network—can become a critical piece of infrastructure for the next generation of AI systems.