AI Research Engineer (Early Career)

Executive Summary

The candidate is an early-career ML practitioner finishing a CS degree in Pakistan with a solid grounding in classical ML and computer vision, and early exposure to LLM APIs. They demonstrate a builder's instinct through multiple completed projects. However, the role at Pergola Studio/AlpacaRelay requires specific, demonstrated competency in LLMs, RAG pipelines, agentic frameworks (LangGraph), vector databases, and prompt engineering — none of which appear meaningfully in the candidate's resume or project portfolio. The absence of code samples, GitHub content, and a professional online presence makes engineering calibration difficult and is a concern for a remote-first, self-directed role. While the candidate shows potential as a developer who could grow into LLM engineering with time and exposure, they do not currently meet the minimum bar for this specific position. They would be a stronger candidate 6-12 months from now if they actively builds LLM/RAG projects and demonstrates them publicly.

Assessment Reasoning

NOT_FIT decision is driven by three compounding factors: (1) Critical skill gap — the role's core technical requirements (RAG, LangGraph, vector databases, AI agents, prompt engineering) are entirely absent from the candidate's demonstrated experience; Python and classical ML alone do not bridge this gap for an applied LLM engineering role. (2) Insufficient evidence of engineering craft — no code sample, no accessible GitHub, and no public portfolio were provided, making it impossible to assess production-code quality, which is an explicit job requirement. (3) Thin professional signal for a remote-first, async, self-directed role — the role requires strong written communication and remote-work independence, yet the candidate submitted no cover letter, no code, and has minimal professional online presence. The candidate is not without merit and shows genuine ML ability, but the gap between current skills and role requirements is too wide for an entry-level position that still demands specific LLM/RAG domain knowledge from day one.

Interview Focus Areas

No code was provided for direct assessment, which itself is a concern for a role that explicitly values production-quality engineering over notebook-style work. Project descriptions suggest functional ML competency at a junior level, but without seeing actual code it is impossible to assess code quality, architectural thinking, or production readiness. The absence of a GitHub profile further limits evaluation.

The candidate presents as a capable early-stage ML practitioner with a traditional ML/CV background and some LLM API exposure. However, the role requires demonstrated hands-on experience with LLMs, RAG, agentic systems, and related tooling (LangGraph, vector DBs, prompt engineering) — none of which are evidenced in the resume. The skill set is adjacent but not sufficiently aligned with the specific applied LLM engineering focus of this position.