3 months ago
Base Salary
$160k - $200k/yr
Responsibilities
- Build and ship production multi-agent systems, including control loops, planning, orchestration graphs, tools, coordination, handoffs, state, and durability.
- Build a safe write-back path for applying agent decisions to customers’ live systems without corrupting data.
- Create evaluation cases, rubrics, and regression tests to measure and improve agent performance.
- Design persistent per-account agent memory, pre-computation, and just-in-time retrieval for fast, reliable reasoning across concurrent instances.
- Protect agents from adversarial data that could hijack their behavior.
- Integrate frontier LLMs behind a model-agnostic abstraction with routing based on task, cost, and latency.
- Own the cost, latency, and reliability of the agent platform surface and establish reusable patterns for faster agent development.
Requirements
- At least 4 years of experience building production systems, including at least 2 years shipping LLM-powered or agentic systems.
- Experience shipping customer-facing LLM or agent systems used by real users, including diagnosing failures and implementing fixes.
- Strong current Python skills; TypeScript or Go is a plus.
- Hands-on experience with modern agent frameworks or orchestration, tool or function calling, structured outputs, retrieval/RAG, and context or memory management.
- Strong systems-engineering fundamentals in concurrency, distributed systems, statefulness, latency, and cost at scale.
- Ability to debug deep, nondeterministic failures in multi-step agent traces.
- High-agency approach to scoping, prioritizing, and shipping work independently.
- Bonus: Salesforce APIs, including Bulk 2.0, Composite, or Pub/Sub, or another large enterprise data source.
- Bonus: MCP, A2A, or similar interoperability standards; Promptfoo, Braintrust, or Langfuse; and Inngest or Temporal.
- Bonus: experience running hundreds or thousands of concurrent agent instances, multi-tenant data isolation with PostgreSQL and RLS, or maintaining a strong customer-data security posture.
- Bonus: founder or founding-engineer experience or contributions to open-source agent or LLM tooling.
Benefits
- Equity is included in addition to base salary.
- The role is based in the Santa Clara, California office.
- Required in-office attendance is Mondays and Wednesdays each week.
Tech Stack
Categories
About LeanData
LeanData builds a SaaS revenue orchestration platform for B2B go-to-market teams, with products for routing, lead-to-account matching, scheduling, and buying-group journeys that run on Salesforce data. It sells subscriptions to enterprises to unify CRM signals and automate customer-lifecycle workflows and analytics across the revenue stack. Founded in 2012 and headquartered in Santa Clara, it counts companies such as OpenAI, NVIDIA, Uber, Snowflake, Anthropic, and DocuSign as customers.