SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salient builds AI agents for regulated financial services, automating loan servicing, compliance, collections, recovery, insurance claims, and disputes for banks and specialty lenders. The company is backed by a16z and Y Combinator with $75M in Series A funding, is cash flow positive with mid-eight figure ARR, and serves over 20% of the auto lending industry while processing millions of real customer calls daily.
You will own the agent product across all channels (voice, text, email, chat), defining what the agent can do, how well it performs, and how quickly new lenders can deploy it. This is a hands-on role requiring you to listen to call recordings, work directly with servicing teams, write detailed specs, and establish quality standards for an agent handling sensitive financial conversations.
Key responsibilities include:
- Roadmap and strategy for the omni-channel agent, with particular focus on seamless channel transitions
- Agent quality metrics (containment, resolution rate, escalation, handle time, CSAT, error/hallucination rates) and evaluation loops
- Servicing workflow catalog, converting bespoke deployments into configurable product features
- Time-to-live optimization for new lender onboarding
- Guardrails and compliance design (FDCPA, TCPA, UDAAP, state servicing rules)
- Standards for discovery, specs, and instrumentation across the product organization
You'll have the option to build a product team over time, but the first year focuses on your own judgment and throughput. The role requires deep technical understanding of LLM-powered systems, ability to reason about prompting, evals, retrieval, tool use, and latency/cost tradeoffs. You must own outcomes rather than artifacts, move fast with incomplete information, and be comfortable with accountability. You'll engage directly with customers and operations teams to understand real workflows, maintain rigorous measurement discipline, and hold your own in technical discussions with engineers. Risk management instincts are essential—designing systems that can fail gracefully in regulated environments.
The team works approximately 60 hours per week, with four days in-person collaboration at the San Francisco office and one remote day.
REQUIREMENTS:
- Shipped LLM-powered products to real users (not demos)
- Can reason about prompting, evals, retrieval, tool use, and latency/cost tradeoffs
- Proven ability to own outcomes and move fast with incomplete information
- Willingness to go to the source—listening to calls and working with ops teams
- Rigorous measurement discipline; ability to distinguish model problems from product problems from data problems
- Technical depth to read traces, debate architecture, and write engineer-respectable specs
- Risk design instincts for regulated environments
- Ability to manage stakeholder relationships, including saying no to large customers
- Clear writing and presentation skills for risk committees
NICE TO HAVE:
- Voice, IVR, contact center, or conversational AI experience
- Background in lending, servicing, collections, or financial operations
- Experience hiring or mentoring product managers
