Applied AI Engineer
About the Company
My client is a fast-growing B2B marketing technology company revolutionising how enterprise teams engage and convert key buyers. Their platform enables hyper-personalised, multi-channel campaigns at scale — empowering marketing and sales teams to identify high-intent accounts, generate tailored content, and execute strategic plays. They are building AI-first features to make this seamless and intelligent, from content generation to account strategy recommendations.
The Role
My client is seeking a senior Applied AI Engineer to lead the development and deployment of novel applications leveraging modern generative AI models.
In this role, you will take the lead in bringing LLM-powered product features to life — building intelligent agents and AI-powered platform capabilities that help marketers drive growth. You'll work closely with product and design teams to prototype, ship, and iterate quickly on AI capabilities.
This is not a role focused on training models from scratch. My client needs an experienced builder who can apply LLMs, APIs, and orchestration frameworks to solve customer problems, build valuable product features, and help shape the AI roadmap.
The Team
You'll join the core product engineering team, working in a fast-moving, collaborative environment where ownership is everything. Everyone contributes to product decisions — from technical tradeoffs to feature planning.
Key Responsibilities
* Build and orchestrate LLM-based agents that plan and execute complex workflows, manage engineering data, and craft custom tools.
* Build workflow infrastructure using Mastra, Vercel AI SDK, or similar TypeScript-based agent frameworks.
* Design agent evaluation strategies using real-world execution data and human-in-the-loop signals.
* Integrate LLM APIs (OpenAI, Anthropic, Hugging Face, etc.) into product workflows and backend services written in TypeScript and Golang.
* Prototype and ship AI-powered product capabilities across the platform — from content generation to account intelligence and campaign automation.
* Design and implement RAG pipelines using vector databases like Qdrant or Weaviate for contextual awareness and retrieval.
* Build robust backend services in TypeScript to serve AI-driven features, including streaming, tool use, and human-in-the-loop workflows.
* Work closely with GTM teams to understand strategic use cases and translate them into intelligent product functionality.
* Continuously explore and evaluate emerging LLM tools, techniques, and architectures.
* Contribute to UX thinking for how AI capabilities are surfaced in the platform.
About You
* 5+ years in software engineering.
* 1–3 years of experience in AI/ML or LLM-based development in a fast-paced startup environment.
* Proven track record building and deploying production-grade agentic systems — ideally automating complex, multi-step workflows.
* Hands-on experience with Mastra, LangChain, LangGraph, Vercel AI SDK, or similar agent orchestration tools.
* Strong understanding of LLM evaluation, prompt design, and domain-specific model tuning techniques.
* Deep understanding of large language models beyond basic API usage, including KV caching, tool integration, and sampling strategies (temperature, top-k, top-p).
* A first-principles mindset and enthusiasm for architecting core AI systems from scratch.
* Full-stack capability — able to own a feature end-to-end, from API to infrastructure to evaluation.
You don't need to meet 100% of the above. More than anything, my client values hunger, passion for this problem space, and a strong track record of achievement.
What's on Offer
* Competitive salary & equity
* 25 days holiday + local public holidays
* Top-spec Apple hardware & remote-friendly setup budget
* London office (Shoreditch) with access to a private gym
* Team offsites twice a year
* Enhanced parental leave
* Regular team socials
Tech Stack
* AI Backend: TypeScript, Mastra, Vercel AI SDK
* Model Integration: Anthropic (Claude), Google Gemini, OpenAI, Perplexity
* Agent Development: Multi-agent orchestration, structured outputs, tool use, streaming, human-in-the-loop workflows
* Cloud: AWS, BigQuery