Applied AI Researcher / Founding Engineer

Executive Summary

The candidate is a strong FIT candidate for the Applied AI Researcher / Founding Engineer role at Pergola Studio. They bring 9+ years of applied ML and LLM experience, elite academic credentials from UCL and Imperial College London, and demonstrated ability to own and scale AI products in production — exactly the profile needed for a founding engineering role. Their experience at Upheal is particularly relevant: they owned the core AI product, reduced costs by 50%, improved quality, and built the eval and observability infrastructure the company relies on. Their founder experience with Lexomat and PayToEat reinforces their comfort with ambiguity, ownership, and cross-functional responsibility. The primary gap to probe is whether their experience extends to low-level model training and fine-tuning (the role's research ambition), or whether it is concentrated at the orchestration and API layer. Pending a successful technical interview, this candidate should advance in the process.

Assessment Reasoning

FIT decision is supported by strong alignment across the majority of required skills: Python engineering, production agentic systems, RAG architectures, prompt engineering, LLM evaluation frameworks (Langfuse), multi-cloud infrastructure, LlamaIndex, tool calling (Claude Agent SDK), and AI observability. The candidate exceeds the experience level bar with 9 years and 7 relevant years, holds elite academic credentials, and brings rare founder experience that is directly valuable at an early-stage startup. Skill gaps in LangGraph, LangSmith, and CrewAI are minor — these are framework-level gaps that an engineer of this caliber can close quickly. The more substantive gap — experience with foundational model training and fine-tuning of distilled models — is a legitimate concern for a role titled 'AI Researcher' but does not disqualify the candidate given the breadth of other requirements they meets. The absence of a code sample is the only notable process gap. Overall, the candidate meets 80%+ of required skills with strong production evidence, warranting a FIT classification with a recommended technical interview to validate the model training depth.

Interview Focus Areas

No code sample or GitHub profile was submitted, making a direct code quality assessment impossible. Based on resume evidence — production system ownership, CI/CD improvements, and diverse tech stack — the candidate likely operates at a Senior level technically, but this must be verified through a technical interview or coding exercise. This is the primary unknown in the evaluation.

The candidate is a highly experienced Senior AI/ML Engineer with 9+ years of end-to-end applied ML and LLM product experience, including strong production credentials at Upheal where they owned the core AI product and delivered measurable cost and quality improvements. Their academic background from UCL and Imperial College London is elite, and their founder experience makes them well-suited for the founding engineer dynamic. The main gap is explicit experience with model fine-tuning/training from scratch and a few specific framework names (LangGraph, CrewAI), though their toolset is otherwise an excellent match.