Software Engineer, AI Agents & Production Systems

ExTrac; Greater London; Hybrid; GBP 65,000 - 90,000; Full time

Benefits offered by this job

Private Medical Insurance

Pension contributions

Parental leave

Nursery scheme

Education budget

33 days leave

Flexible working

London office

Job summary

ExTrac is seeking a Software Engineer to join our AI team, building Co-Analyst and the analytical AI features around it. You will own well-defined features end to end in a Python and Go codebase, collaborating with senior engineers, the data team, and an ML-focused group who train and evaluate models we use.

You will work on agent orchestration, retrieval, graph analytics, and long-running streaming pipelines, shipping production services and APIs with latency, cost, and reliability in mind.

Qualifications

• 2+ years of professional software engineering experience
• Proficiency in building services in Python, with working knowledge of Go

Responsibilities

• Build and improve components of the agent loop: context assembly, tool selection, sub-agent orchestration.
• Develop analytical AI features from network construction to summaries analysts read.
• Run experiments against live analyst traffic behind feature flags.
• Contribute to agreeing what 'better' means and flag ineffective approaches.
• Work with embeddings beyond retrieval for network and community tasks.
• Operate in a model-agnostic design, swapping models and embeddings as needed.

Skills

Python

Go

Distributed systems

Cloud infrastructure

CI/CD pipelines

Documentation

Tools

Elastic

Job description

ExTrac is seeking a Software Engineer to join our AI team, building Co-Analyst and the analytical AI features around it. You will own well-defined features end to end in a Python and Go codebase, collaborating with senior engineers, the data team, and an ML-focused group who train and evaluate models we use.

You will work on agent orchestration, retrieval, graph analytics, and long-running streaming pipelines, shipping production services and APIs with latency, cost, and reliability in mind.