The role in one sentence
You are a senior, AI-native delivery lead who owns engagement outcomes end to end: shaping the solution, defining and planning the work, leading the client, and making sure what we build is adopted and delivers measurable value.
Why this role exists
We're building a small, senior delivery team that delivers what much larger teams traditionally would, by using AI across the full surface of the work.
AI has made building faster. It hasn't made it any easier to know what to build, get the right decisions made, or get people to adopt what's been built. That's where engagements succeed or fail, and it's what this role owns: making sure the right thing gets built, it lands, and the client's team comes out the other side working differently.
You partner with a Technical Lead / AI-Forward Engineer, working as one team: you bring the problem, the outcome, the sequencing, and the adoption; the TL brings the architecture and the build; both of you own whether it works.
This is a senior individual contributor role: you lead engagement teams, not a reporting line. Teams are small (typically 2-4 people), you run one or two engagements at a time, and clients range across mid-market industries. Engagements run from weeks-long plan-and-define and POC work to multi-year managed services. Fastloop's own delivery is the first proving ground: we are becoming AI-native at the same time as the clients we serve.
What you'll own
- Taste and judgment. Plausible plans, backlogs, and status reports are now cheap to produce. Knowing whether a story is ready, a plan is real, "on track" is true, and the work solves the problem is not. Your judgment on what's ready, done, and at risk is the bar.
- Solution shaping. You lead discovery with the TL and shape the solution together: the problem worth solving, the outcome worth paying for, what comes first, and what the client can absorb. You challenge the ask when there's a better way, including when that way isn't AI.
- Workflow redesign. You map how work actually flows across people and systems and redesign it: where agents act, where deterministic automation fits, and where a human decides.
- Documentation and breakdown. You own how work is specified: requirements, stories, and acceptance criteria clear enough for an agent or engineer to execute without a meeting, sliced vertically, with testable definitions of done.
- Planning and visibility. You plan for AI-speed delivery, where the constraints are review, validation, decision latency, and adoption rather than build time. Status comes from the source of truth, not a hand-built deck, and you re-plan from evidence.
- Team transformation. You change how client teams document, break down, plan, and report work, connecting AI into their tracker, source control, and CI/CD, and growing adoption beyond the pilot team.
- Quality and governance. You define ready and done, what "good" means for AI outputs in business terms, where humans must approve, and how sensitive data is handled.
- Outcomes and value. You agree the business outcome up front (cost, cycle time, quality, throughput, adoption), measure it, and report against it.
- Engagement and account. You own scope, risk, dependencies, resourcing, margin, and delivery health. You lead client conversations from working sessions to steering committees, grow accounts through the quality of delivery and a roadmap clients want to fund, and support pre-sales with discovery, estimates, and SOWs.
- Delivery library. You grow the team's shared library of delivery skills, agents, templates, and playbooks, and with the engagement sponsor turn proven patterns into Fastloop offerings.
How you work: the compounding loop
This is how we expect the work above to get done. You don't do it by hand: you build and run the machine that does it, own the judgment on what comes out, and contribute back so it gets better. That's the difference between an orchestrator and an operator who uses AI.
Specs, breakdowns, plans, forecasts, and status are produced by skills and agents you build and run. Your time goes to the judgment calls, the client, and making the machine better. The engagement you land matters; so does what Fastloop and the client can do next time that they couldn't before.
- Start from golden paths. Begin every engagement from the team's existing skills and playbooks (story development, scoping, status, discovery) and improve them. Rebuilding a template the team already has creates drag for everyone who comes after you.
- Contribute back continuously. A better story format, a sharper estimation heuristic, a status agent that finally reads CI correctly: it goes back into the shared system quickly rather than staying trapped in one client's workspace.
- Instrument delivery itself. Treat delivery as a system you can measure. Cycle time, decision latency, rework, forecast accuracy, and adoption are tracked, not guessed, so you can tell whether a change to how we work actually helped.
- Protect machine-maintenance time. A meaningful slice of every cycle goes to hardening skills, refining playbooks, documenting decisions, and improving the delivery harness. That investment is what makes the next engagement start further ahead.
- Commit to the AI-native way. AI is the default operating model for delivery here, not a side experiment running next to the old way. The toolset has a short half-life, so staying current with new model, agent, and harness capabilities is expected. So is the judgment to know when an agent, a deterministic automation, or a human conversation is the right tool.
If your instinct is to land the engagement in front of you and improve the system that lands the next one, you'll fit.
What you bring
Delivery: (required)
- Significant experience leading software, data/ AI engagements from ambiguity into production, owning the outcome rather than a workstream.
- Real experience building and running agent workflows for your own delivery work (Claude, Claude Code, Cursor, or equivalents), not just a chat window you paste into.
- Hands-on enough to prototype a workflow, draft a skill, or inspect agent outputs and traces, and to know when a plan, story, or output is wrong.
- Strong, opinionated craft in requirements, story writing, and backlog design.
- Business analysis, process mapping, and workflow redesign, with working knowledge of enterprise systems (ERP, CRM, asset management, finance, etc) and modern data and analytics.
- Working fluency with modern software delivery: Git, code review, Jira or Linear, CI/CD.
- Enough understanding of agentic systems to plan and govern them: context engineering, MCP and tool integration, evals, human-in-the-loop design, permissions, and failure modes.
- A track record of changing how teams work, with adoption that stuck.
- Client leadership from delivery teams to executives, and commercial judgment across scope, estimates, SOWs, and engagement economics.
- Depth in a relevant domain, such as data and analytics or AI-native software delivery, is a strong plus.
Google Cloud and Anthropic ecosystems
You can hold a credible solution conversation across both:
- Google Cloud: BigQuery, Looker, Vertex AI and Gemini, Cloud Run, Pub/Sub, Cloud Composer.
- Anthropic: Claude models, the Claude Developer Platform, and Claude Code, including enterprise deployment: permissions, security, governance, and evaluation.
Experience with Databricks, Snowflake, Azure, or AWS is valuable. A Google Cloud Professional certification or Claude Certified Architect credential is a plus.
How you operate
- Ownership. You own the engagement outcome end to end, including dependencies, decisions, and constraints outside your immediate remit.
- Comfort with ambiguity. You take a vague brief, find the underlying problem and the measurable outcome, and produce a clear path: priorities, owners, next steps.
- Bias to ship. You get a first vertical slice into the real environment early and plan from reality rather than polishing a roadmap in isolation.
- Product sense. You care whether the system actually solves the problem. You understand the people and workflows around it, and you build adoption in from the start.
- Judgment over dogma. Scrum, SAFe, and Prosci are tools, not religions. You apply the right amount of structure, and you know the difference between automation that compounds and process that just adds ceremony.
- One team. You work with the TL as one team, create clarity without becoming the bottleneck, and challenge the plan, the client, or your partners when the evidence says to.
- Grow the people around you. Client teams and Fastloop colleagues develop stronger judgment and delivery craft from working with you.
- Compounding instinct. You naturally turn repeated delivery work into better skills, sharper templates, clearer interfaces, and playbooks the rest of the team can build on.
You'll know you're succeeding when
- Clients understand the outcome and why it matters, teams know what good looks like, decisions happen on time, and risks surface early.
- Solutions reach production, are adopted, and hold a measurable improvement after we step back.
- Specs, plans, and status come from the machine, and your time goes to judgment, the client, and improving it.
- Client teams work differently, and the machine you installed keeps running without us.
- Each engagement starts further ahead than the last because you contributed back to the library.
- Stakeholders trust your judgment, including when you challenge the plan, and accounts grow because of the work.
What this is not
- A project administration seat where you chase updates and assemble statuses by hand.
- A role where success ends with a roadmap, a requirements document, or a signed SOW.
- A role that operates at arm's length from the technology. You don't need to be the primary engineer, but you need to read the work.
- A traditional delivery manager role. Your value isn't measured by how many ceremonies you run or reports you write. It's measured by whether the work lands, whether teams work better afterwards, and whether the next engagement starts further ahead.
We're looking for delivery leaders who can go broad without becoming shallow, use AI without outsourcing judgment, and leave every team they touch working better than they found it.
If that already sounds like how you work, you'll probably feel at home here.
