Full-time
MTS - Agent Platform
San Francisco, California, United States · On-site
$220,000 - $320,000
ML Engineer, Agent Platform
Getting an AI agent into production for a new enterprise customer usually means an engineer spending weeks turning that customer’s processes into something that actually works. That doesn’t scale. This team is building the system that does that job instead: it takes a customer’s procedures, generates the agents, tests them and improves them, so new customers can go live in days and eventually set themselves up.
They’re an early-stage company building AI agents for operational work where mistakes are expensive. Their agents handle long, multi-step tasks across voice, text, email and documents. The founders have built and sold several companies before, the engineering team includes multiple former CTOs, and they already have paying enterprise customers.
At its heart this is a specification problem. Each agent comes from a structured definition of how a customer works: its rules, conditions and escalation paths. You’d help design what that definition can express and make sure the agents built from it behave as intended. The work covers:
Generating agents from a single versioned definition, plus the developer loop around it, including replaying real conversations
Evals, user simulation and verification that prove an agent is ready before it goes live
Safe deployment: staged rollouts, review steps, rollback and human hand-off
Tracing and monitoring across a fleet of live agents
Agents that can spot their own failures and improve
Customer-facing onboarding and configuration tools
You’ll have shipped production software that real users depend on, and ideally built something like an agent builder, workflow engine or developer platform. You know when deterministic code beats a model and when it doesn’t, and how to test systems that don’t behave the same way twice. A grounding in logic, statistics or formal languages is a strong plus, as is experience with DSLs, compilers or formal verification.
The team is small with no layers of management, so you’d own features from idea to customer launch.
📍 San Francisco, 4 days onsite. Visa support available
💰 Up to circa $320K base + stock, healthcare and unlimited PTO
✅ Repeat founders with multiple exits | Paying enterprise customers
If it sounds relevant, we’d be happy to talk it through.
All applicants will receive a response.
Salary
$220,000 - $320,000
Location
San Francisco, California, United States · On-site
Contract
Full-time
Your consultant
Kieran Newton

Full-time
MTS - Agent Platform
San Francisco, California, United States · On-site
$220,000 - $320,000
ML Engineer, Agent Platform
Getting an AI agent into production for a new enterprise customer usually means an engineer spending weeks turning that customer’s processes into something that actually works. That doesn’t scale. This team is building the system that does that job instead: it takes a customer’s procedures, generates the agents, tests them and improves them, so new customers can go live in days and eventually set themselves up.
They’re an early-stage company building AI agents for operational work where mistakes are expensive. Their agents handle long, multi-step tasks across voice, text, email and documents. The founders have built and sold several companies before, the engineering team includes multiple former CTOs, and they already have paying enterprise customers.
At its heart this is a specification problem. Each agent comes from a structured definition of how a customer works: its rules, conditions and escalation paths. You’d help design what that definition can express and make sure the agents built from it behave as intended. The work covers:
Generating agents from a single versioned definition, plus the developer loop around it, including replaying real conversations
Evals, user simulation and verification that prove an agent is ready before it goes live
Safe deployment: staged rollouts, review steps, rollback and human hand-off
Tracing and monitoring across a fleet of live agents
Agents that can spot their own failures and improve
Customer-facing onboarding and configuration tools
You’ll have shipped production software that real users depend on, and ideally built something like an agent builder, workflow engine or developer platform. You know when deterministic code beats a model and when it doesn’t, and how to test systems that don’t behave the same way twice. A grounding in logic, statistics or formal languages is a strong plus, as is experience with DSLs, compilers or formal verification.
The team is small with no layers of management, so you’d own features from idea to customer launch.
📍 San Francisco, 4 days onsite. Visa support available
💰 Up to circa $320K base + stock, healthcare and unlimited PTO
✅ Repeat founders with multiple exits | Paying enterprise customers
If it sounds relevant, we’d be happy to talk it through.
All applicants will receive a response.
Your consultant
Kieran Newton

Full-time
MTS - Agent Platform
San Francisco, California, United States · On-site
$220,000 - $320,000
ML Engineer, Agent Platform
Getting an AI agent into production for a new enterprise customer usually means an engineer spending weeks turning that customer’s processes into something that actually works. That doesn’t scale. This team is building the system that does that job instead: it takes a customer’s procedures, generates the agents, tests them and improves them, so new customers can go live in days and eventually set themselves up.
They’re an early-stage company building AI agents for operational work where mistakes are expensive. Their agents handle long, multi-step tasks across voice, text, email and documents. The founders have built and sold several companies before, the engineering team includes multiple former CTOs, and they already have paying enterprise customers.
At its heart this is a specification problem. Each agent comes from a structured definition of how a customer works: its rules, conditions and escalation paths. You’d help design what that definition can express and make sure the agents built from it behave as intended. The work covers:
Generating agents from a single versioned definition, plus the developer loop around it, including replaying real conversations
Evals, user simulation and verification that prove an agent is ready before it goes live
Safe deployment: staged rollouts, review steps, rollback and human hand-off
Tracing and monitoring across a fleet of live agents
Agents that can spot their own failures and improve
Customer-facing onboarding and configuration tools
You’ll have shipped production software that real users depend on, and ideally built something like an agent builder, workflow engine or developer platform. You know when deterministic code beats a model and when it doesn’t, and how to test systems that don’t behave the same way twice. A grounding in logic, statistics or formal languages is a strong plus, as is experience with DSLs, compilers or formal verification.
The team is small with no layers of management, so you’d own features from idea to customer launch.
📍 San Francisco, 4 days onsite. Visa support available
💰 Up to circa $320K base + stock, healthcare and unlimited PTO
✅ Repeat founders with multiple exits | Paying enterprise customers
If it sounds relevant, we’d be happy to talk it through.
All applicants will receive a response.
Your consultant
Kieran Newton


