Agentic Workflow Lead
Summary
Loom is hiring a senior practitioner to lead delivery of agentic AI engagements. The role exists because the Applied AI and Harness Frameworks services need a dedicated lead to run concurrent customer engagements without principal bottlenecks.
What you’ll do
- Lead delivery of multi-agent systems for enterprise customers, from scoping through production.
- Architect agent runtimes, tool integrations, and evaluation pipelines for production-grade systems.
- Run forward-deployed engagements: embedded with the customer team, shipping in their codebase against their real infrastructure.
- Set Loom’s technical standard for how we build agentic systems — patterns, eval frameworks, safety guardrails.
- Partner with the principals on architectural decisions and mentor practitioners across the team.
What we’re looking for
- Senior-level experience building multi-agent systems in production environments.
- Production track record with at least one agent framework (LangGraph, CrewAI, AutoGen, or a proprietary equivalent).
- Direct experience operating LLM-based systems at scale: evaluation, guardrails, cost management, latency.
- Comfort working in a customer’s codebase against their real constraints.
- Strong technical communication. You can explain trade-offs to engineers and to executives in the same conversation.
What is helpful but not required
- Prior experience at a forward-deployed company.
- Background in distributed systems or inference infrastructure.
- Open-source contributions to agent frameworks or evaluation tooling.
What we offer
- Competitive base compensation, leveled against senior practitioner roles at top AI companies.
- Meaningful equity participation. Compensation is transparent and level-justified, with a written rationale.
- Remote-friendly across the US, Canada, Mexico, and India.
- Direct work with the principals and direct accountability for engagement outcomes.
- Hiring process: four stages, decision within ten business days of the working session. See the full process below.
Hiring Process
Designed around what predicts performance.
Four stages, structured the same way for every candidate. The hiring decision rests on a work sample scored independently by two evaluators — the methods with the strongest evidence behind them in the research on hiring. A decision within ten business days of the work sample.
Structured conversation
Thirty minutes with Mugdha or Kevin. The same set of questions for every candidate at this stage: your recent work, the problems you want to be solving next, and what would make this engagement work for you. Mutual-fit calibration. The hiring decision is not made here.
The hiring decision
A scoped problem drawn from a current Loom engagement — the same prompt every candidate for the role has seen. Two hours, paired with one of us. We score independently on a written rubric (technical depth, judgment under uncertainty, collaboration, communication) and compare scores before discussion.
Structured interviews
Two or three thirty-minute conversations with the practitioners you would work with. Each is structured around a small set of behavioral and situational questions. Panelists score independently and calibrate after. You are evaluating us in the same conversations.
Transparent and timely
A decision within ten business days of the work sample. If we extend an offer, you receive a written rationale for the level, the comp band, and the equity philosophy. If we do not, you receive specific feedback from the work-sample rubric.
Process design based on the meta-analytic evidence on selection-method validity: Schmidt & Hunter (1998), reanalyzed by Sackett, Zhang, Berry, & Lievens (2022). Work samples and structured scoring with multiple independent raters consistently rank among the highest-validity selection methods. Unstructured interviews, brainteasers, and single-evaluator judgments rank substantially lower.
Apply
Submit your resume and tell us what about this role interests you. ● indicates a required field.
Every application is read by Mugdha Pandit or Kevin Dickerson — no ATS in between.
Application received
Thank you, there.
Your application for this role is in. We received your resume and we'll review it ourselves — every application gets read by Mugdha Pandit or Kevin Dickerson.
What happens next. If your background is a strong fit, you'll hear from us within five business days with next steps. If we don't see a fit for this role right now, you'll still hear back — we won't leave you in silence.
Reference
—
Need to reach us directly? Email careers@loom.technology and quote that reference.