Software Engineer, Agent

Written by the T&S Network from Variance's posting.

Building AI agents that make decisions inside fraud, risk investigation, and identity review workflows, so enterprises spend less time on repetitive manual reviews. Sits in the technical delivery team working directly with customers from first demo through production and expansion, including marketplaces, gig platforms, fintechs, and financial institutions.

• Fraud and scams

What you'd do

• Build AI agents that decide cases in fraud and risk pipelines, sized for millions of investigations in each deployment.
• Design evaluations and proofs of concept from each customer's data and pain points, and run pilots and production deployments.
• Translate customer edge cases into technical evaluations and serve as the technical voice with executive stakeholders.
• Sit with operational stakeholders and encode their investigative judgment directly into agents.
• Take agents through scoping, context and tool engineering, decision-quality evals, and tuning on production data.
• Set up customized demo environments on customer data during pre-sales technical discovery with the sales team.

What it asks for

• Three to five years in customer-facing engineering, whether forward-deployed, solutions, or software work with direct client contact.
• Strong software engineering fundamentals and a record of shipping end-to-end production systems.
• Practical work building LLM applications, covering prompt and context design, tool calling, and evaluation.
• Ability to turn ambiguous customer workflows into working systems.
• Skill at briefing both engineering teams and executives on complicated technical topics.
• Experience deploying agentic systems in production (preferred).
• Domain experience in fraud or risk (preferred).

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Variancenorth_east before you apply. If we got something wrong, report this listing.