Skip to content

Variance

Software Engineer, Agent

Location
San Francisco, CA, US
Arrangement
On-site
Employment type
Full-time
Posted
25 August 2026 (about a month ago)

Checked about a month agoApplications go to the employer, never to RoleSprint

About this role

At Variance, we're teaching machines to make the hardest judgement calls at scale. That means building AI agents for the high-precision gray-area of risk investigations, fraud, and identity reviews. Your work will directly help enterprises spend less time on repetitive manual reviews and investigations, allowing them to focus on what matters most. We’re rapidly growing, just raised a $21M Series A, and are building a team that works directly with leaders from leading marketplaces, gig platforms, fintechs, and financial institutions.

Why this role

In this role, you'll own the technical delivery of Variance's agents for enterprises operating in complex, high-consequence environments. You'll build agents that make real decisions inside fraud and risk workflows, working directly with customer teams from the first demo through production and expansion.

We're seeking someone with the creativity to push the boundaries of tasks previously thought impossible with AI agents, and advance our mission of protecting people and our institutions. This is a foundational hire for this position and you’ll have the opportunity to build and shape the future of this function.

This is an in-person position. Our office is located in San Francisco, CA.

What you'll do

- Design and ship production-grade AI agents. You’ll build agents that make decisions inside fraud and risk workflows, partnering directly with our customers at the scale of millions of investigations per deployment.

- Own the technical evaluation. You'll design evaluations and PoCs based on the customer’s data and pain points, and drive the technical execution of pilots and production deployments. You will translate customer edge cases into the technical evaluation, and you will be the technical voice in a room of executive stakeholders.

- Learn how the work is done: You'll sit with operational stakeholders, and directly encode their judgement into the agent.

- Own the agent development lifecycle end to end. You’ll take agents from the initial pilot through deployment and continuous iteration by scoping the workflow with the customer, engineering the context and tools the agent needs, building evals that actually measure decision quality, and tuning performance against real production data. You set the bar for what a well-built agent looks like.

- Design customized demos. You'll be embedded in the pre-sales process and partner with our sales team to run technical discovery. You'll then help showcase the power of Variance on a customer's own data by setting up demo environments tailored to their industry and key use case.

- Shape the roadmap long-term. The patterns you see across deployments drive what Variance builds: the context engine, entity graph, and agentic review automation. What one customer needs today may define the roadmap for every customer in that industry.

Example projects

- Build the onboarding verification agent for an enterprise marketplace, helping to promote and support the expansion into international geographies.

- Design the supervisor layer that verifies agent decisions before they reach an investigator's queue

- Engineering the eval harness that helps align the detection of fraud against millions of samples of real-world fraud.

What you'll bring

- 3-5 years in customer-facing technical roles, forward-deployed or solutions engineering, or software engineering with direct customer exposure

- Strong software engineering fundamentals with a track record shipping end-to-end production systems

- Hands-on experience building with LLMs, including prompt and context engineering, tool use, and evals, plus calibrated instincts about what models can and cannot do reliably

- Strong problem-solving in ambiguous, fast-changing environments. You can take a fuzzy customer workflow and turn it into a working system.

- Comfort working directly with customers with the ability to communicate complex technical concepts to both engineers and executives

- Bonus if you have familiarity with deploying agentic systems in production, domain experience in fraud or risk, or experience shipping into regulated environments (security reviews, data residency, audits)

What we offer

- Health care: platinum-level medical, dental, and vision insurance

- Unlimited PTO, sick leave, and parental leave

- Up to $100 per month in reimbursement for personal health and wellness expenses

- Up to $200 per month in commuter benefits

- Free lunch and dinner

- 401(k) plan with employer match (eligible after 6 months)

- Meaningful equity in a fast-growing, Series A company

We evaluate all applicants consistently without regard to race, color, religion, gender, national origin, age, disability, veteran status, pregnancy, gender expression or identity, sexual orientation, citizenship, or any other legally protected class. We encourage you to apply even if your experience doesn't precisely match the job description.

Work location

  • San Francisco, CA, US

Related jobs

Ready to make a decision?

This role is either worth your time or it isn’t.

Analyze the posting against your experience, see the gaps clearly, and build the right materials only if the opportunity makes sense.

Nothing is submitted automatically. You choose what happens next.

About this listing

Published on Ashby under the board identifier Intrinsic Safety, which is the name the employer’s own job board carries. RoleSprint has not verified the company’s registered or trading name, so it is shown exactly as published rather than tidied up.

RoleSprint is not the employer and not a recruiter. Applications are made on the employer’s own site and never reach us; what RoleSprint does is help you decide whether a role is worth your time and prepare for it if it is.

Published 25 August 2026, last checked about a month ago. A posting stops being advertised here 90 days after the employer published it, and one the employer takes down is marked closed rather than quietly removed.

More searches like this one

Browse all current openings

No credit card requiredStart free