Founding AI Engineer (Part Time) - Prompt Engineering & LLM Applications

Part-Time Founding AI Engineer - Prompt Engineering & LLM Applications About Y22 AI Y22 AI is building an AI sales training and call intelligence platform that helps sales teams identify what their top reps are doing differently, score real sales calls, and turn rep weaknesses into targeted AI roleplay training. Our platform helps reps practice against AI prospects that mirror the exact buyers, titles, companies, objections, and sales situations they face in the real world. We are looking for a part-time Founding AI Engineer with strong prompt engineering and LLM application development experience to help build the core intelligence layer behind Y22. This is an early-stage, equity-based role for the first 6 months. We are looking for someone who wants meaningful ownership, can move fast, and is excited to help build a category-defining AI sales coaching platform from the ground up. What You'll Work On Create realistic AI sales prospects through advanced prompt engineering Tune AI prospect behavior by buyer type, title, industry, company size, pain points, objections, call type, and training goal Build prompt systems that make roleplays feel realistic, challenging, adaptive, and useful for sales reps Turn real sales calls into structured coaching insights Score sales conversations using Y22's sales performance rubric Analyze transcripts for rep strengths, weaknesses, objections, talk tracks, missed opportunities, and next steps Generate personalized roleplay scenarios based on each rep's performance gaps Build prompt templates, prompt chains, evaluation workflows, and structured output systems Create RAG pipelines over customer sales calls, product docs, ICP docs, sales playbooks, objection libraries, and training materials Improve AI accuracy, consistency, relevance, hallucination resistance, and output quality Support future real-time sales assistant workflows Responsibilities Design, test, and iterate prompts for AI roleplay, call scoring, coaching insights, and sales simulations Build prompt systems that control tone, difficulty, objection patterns, buyer resistance, personality, and conversation flow Develop LLM-powered backend workflows using Python Create structured outputs for scorecards, coaching plans, roleplay scenarios, and rep performance insights Build transcript analysis and scoring workflows grounded in call evidence Design retrieval workflows using embeddings, vector databases, and customer-specific knowledge bases Develop evaluation processes to measure prompt quality, hallucination risk, relevance, realism, and consistency Monitor model outputs and improve prompts based on real user behavior and feedback Work closely with the founder on product direction, customer feedback, and roadmap priorities Move quickly while maintaining strong engineering standards Required Qualifications Strong prompt engineering experience with LLMs Strong Python experience Experience building LLM-powered applications Experience with OpenAI, Anthropic, Grok, or similar model providers Ability to design prompts for structured outputs, roleplay behavior, scoring, classification, summarization, and coaching Understanding of prompt testing, prompt versioning, temperature testing, and output evaluation Experience with RAG, embeddings, retrieval, and vector databases SQL/database experience Backend/API development experience Docker experience Working knowledge of AWS or GCP Ability to evaluate AI outputs for accuracy, relevance, hallucination risk, and usefulness Strong product instincts and ability to operate in an early-stage environment Comfortable working part-time in a fast-moving startup Nice to Have Experience building AI roleplay, simulation, tutoring, coaching, or conversational training systems Experience designing AI personas, buyer behavior, objection handling, or adaptive conversation flows Experience with LangChain, LangGraph, LlamaIndex, or similar frameworks Experience with voice AI or real-time conversation systems Experience with prompt injection prevention and AI safety guardrails Experience with AI evaluation, model monitoring, and output logging Experience with fine-tuning, LoRA, PEFT, or model optimization Kubernetes or production infrastructure experience Experience with sales tech, call intelligence, CRM integrations, coaching tools, or enablement platforms Who You Are You are excellent at