What Is a Pod?
A pod is a self-contained engineering unit designed to operate independently inside a client environment. Think of it as a special ops team for software: small, focused, and mission-driven.
Every pod has the same core structure, regardless of the engagement type.
Pod Composition
The Pod Lead
- Who: A senior engineer with 8+ years of experience and at least 3 FDE deployments under their belt.
- Role: Technical decision-maker, client relationship owner, quality gatekeeper.
- Accountability: Pod leads own the outcome. If the project fails, it's on them.
The Engineers (2-3)
- Who: Mid-to-senior engineers with deep domain expertise matched to the engagement.
- Role: Write code, build integrations, deploy systems. The hands on the keyboard.
- Selection: Shadow-matched to the client's tech stack and problem domain.
The Data Specialist (LLM Foundry only)
- Who: A data engineer or ML engineer with experience in data curation and model training.
- Role: Manages the data pipeline, quality metrics, and training infrastructure.
Daily Rituals
Morning Standup (15 min)
Every pod runs a standup at the start of the client's business day. Not our business day—theirs. If the client is in Tokyo, standup is at 9 AM JST.
Format:
- What shipped yesterday
- What's shipping today
- Blockers (if any)
Red Phone Check (continuous)
Every pod has a "red phone"—a direct Slack channel or phone number that the client champion can use 24/7. When the red phone rings, someone answers. Always.
Weekly Demo (30 min)
Every Friday, the pod demos what shipped that week to the client stakeholder. Not a status update. Not a slide deck. A live demo in production.
Retrospective (bi-weekly)
Every two weeks, the pod runs an internal retro:
- What went well?
- What didn't?
- What do we change next sprint?
How We Measure Success
Every pod tracks three metrics:
1. Velocity
How fast are we shipping? Measured in story points per week, but more importantly, in features deployed to production.
2. Quality
What's our defect rate? We target zero production incidents per sprint. When they happen (and they do), we run a blameless post-mortem within 24 hours.
3. Client Satisfaction
We run a monthly NPS survey with the client stakeholder. Our target is 9+/10. Our actual average across all pods is 9.4.
Communication Protocols
With the Client
- Slack/Teams: Real-time communication in the client's workspace
- Weekly email summary: Sent every Friday after the demo
- Monthly business review: Pod lead + client leadership, focused on outcomes and roadmap
With FDE Team HQ
- Daily pod report: Automated summary of commits, PRs, and deployments
- Weekly pod lead sync: All pod leads meet to share learnings and patterns
- Escalation path: Pod lead → Engineering Director → CEO. Three steps max.
Pod Types
Integration Pod
Focused on connecting the client's product to enterprise systems. Salesforce, SAP, Oracle, Workday, custom APIs. The bread and butter of FDE work.
LLM Foundry Pod
Focused on building and deploying domain-specific AI models. Data curation, fine-tuning, evaluation, and production deployment.
Special Ops Pod
For everything else. Security audits, performance optimization, architecture redesign, technical debt reduction. The Swiss Army knife.
The Bottom Line
A pod isn't a traditional outsourcing arrangement. It's a deployment methodology designed to deliver outcomes in the fastest possible timeframe. Every role, every ritual, every metric is optimized for one thing: shipping value to the client. It's a deployment methodology designed to deliver outcomes in the fastest possible timeframe. Every role, every ritual, every metric is optimized for one thing: shipping value to the client.
Want to see how a pod would work for your project? Book a discovery call and we'll design a custom pod for your needs.