Applied AI Engineer - Agentic

Applied AI Engineer - Agentic AI

Key Details

• Location: Reading
• Compensation: Competitive salary + meaningful equity
• Stage: Early-stage, venture-backed AI startup
• Team: Working directly with founders and early engineering team

Overview

We’re building autonomous AI agents that perceive, reason, plan, and act across live environments.

This is not a prompt-engineering role.

You’ll build production-grade agent systems capable of navigating interfaces, using tools, orchestrating workflows, recovering from failure, and operating reliably at scale.

We’re looking for engineers who can bridge frontier AI research with hardened distributed systems.

You’ll work directly with founders to define the architecture, infrastructure, and operational backbone behind next-generation AI agents.

Key Responsibilities

• Build autonomous AI agents using modern agent frameworks such as LangGraph, OpenAI Agents SDK, AutoGen, CrewAI, or similar orchestration tooling
• Design multi-agent systems with planning, memory, tool use, routing, reflection, and long-running execution
• Develop agents powered by frontier LLMs and VLMs
• Build retrieval pipelines using vector databases such as Pinecone, Weaviate, Chroma, pgvector, or Milvus
• Engineer production AI infrastructure using Python, FastAPI, TypeScript, gRPC, event-driven systems, and async architectures
• Deploy and scale agent systems using Docker, Kubernetes, Helm, Terraform, AWS/GCP, and modern CI/CD pipelines
• Build observability and evaluation systems using LangSmith, Weights & Biases, MLflow, OpenTelemetry, Grafana, or custom telemetry pipelines
• Own model serving, inference optimisation, batching, caching, and GPU orchestration using tools like vLLM, Ray Serve, Triton, or BentoML
• Develop feedback and evaluation loops using RLHF, SFT, synthetic data generation, automated evals, and human-in-the-loop systems
• Build agents that interact with APIs, browsers, game engines, desktop environments, or mobile systems using Playwright, Selenium, browser automation, or custom tooling
• Solve hard production problems around hallucinations, memory drift, latency, concurrency, recovery, and agent reliability
• Operate with high ownership in a fast-moving startup environment with ambiguity and technical depth

Next Steps

If this sounds like your kind of challenge, we’d love to hear from you.

The process is designed to move quickly and give you real exposure to the team, technical problems, and opportunity:

• Introductory conversation with the hiring managers / founders
• Collaborative Conversation around product, scaling, and engineering approach