How Teams Are Staffed
Client engagements are the course. This page is how you end up on one, and how you influence which one.
Two things decide it. Fit determines which projects you are eligible for. Points determine which eligible project you get. Fit is a gate, not a tiebreaker, and it always runs first.
The Pod
One engagement, one pod. A pod is the smallest group that can find the problem, build the thing, and hand it over without a handoff inside the team.
| Role | Who | Hours |
|---|---|---|
| Engagement lead | MBA | 4 per week |
| Builders (two) | Undergraduate builders or technical MBAs | 10 per week each |
| QC lead | The Product Evaluation and QC function | 2 per week |
| Design lead | The UX/UI function, on demand | As needed |
| Faculty advisor | Scott Murff | Oversight |
An engagement’s length is set with each client; five weeks is common. A team is one or two builders, 10 hours each per week. QC, design, and faculty time ride on top of that. This is why the one rule that never bends is that a builder is on exactly one engagement at a time. Your ten hours are sold.
Every stint starts with a tentative roll-off date agreed with the client. At that point the team either extends with the client or rolls off and is available for the next engagement.
Engagement lead owns the client relationship: the weekly call, the weekly update, the decision log, scope, and escalation. Owns the answer to “does this still fit in the time the client agreed to.”
Builders build. There are no builder roles: both write code, both own the architecture, and both are accountable for shipping.
QC lead is a gate. Nothing reaches the client without passing it, and the review happens at kickoff and at every phase boundary, not the night before a milestone.
An engagement lead can carry two engagements. A QC lead can carry four. Builders carry one.
The Six Builder Archetypes
Every project gets classified at intake with a primary and a secondary archetype. Every builder carries a self-reported rating on all six, calibrated by faculty. This is the matching key.
| Archetype | What it looks like in practice | |
|---|---|---|
| A | Full-stack application | Net-new web or mobile app. Auth, database, deploy |
| B | AI and agents | LLM pipelines, scoring, evals, retrieval, agent orchestration |
| C | Data and modeling | Forecasting, optimization, analytics, decision frameworks |
| D | Integration and automation | APIs, webhooks, document processing, workflow glue |
| E | Product design | Design-heavy surfaces, design systems, net-new interfaces |
| F | Domain and strategy | Regulated industries, market analysis, business rules |
Ratings run 0 to 3. 0 none. 1 coursework only. 2 you have shipped it once. 3 you can lead it and teach it.
These map onto the Builder hexagon from the program overview. The hexagon is about your development. The archetypes are about what a specific client is paying for this semester.
The Matching Rule
A pod is validly staffed only when all four are true:
- At least one builder is rated 3 in the project’s primary archetype.
- The two builders together cover the secondary archetype at 2 or better.
- Neither builder is on another active engagement.
- Someone else on the bench, outside the pod, is also rated 3 in the primary archetype.
Rule 4 is the two-deep rule. It exists so that one person dropping the course does not end a client engagement. If we cannot satisfy rules 1 and 4, we do not sign the build phase. We sell discovery only, or we decline the project.
That is a real constraint on the kind of work the Foundry can take, and it is the honest reason some projects get turned down.
Points
You get 100 points per month, capped at 300. Points buy one thing: a seat on the project you want.
| Accrual | 100 points per month, every enrolled builder |
| Cap | 300. Accrual stops there |
| While staffed on an engagement | Points keep accruing, still capped at 300 |
| When you win a seat | Balance resets to zero |
| Carryover between semesters | Yes, up to the cap |
| Any other use | None |
The cap and the reset exist so the market is not decided by whoever has been idle longest. Sitting on the bench since September should be an advantage. It should not be an unbeatable one.
The Auction
- A new engagement signs. We post the project brief: what it is, primary and secondary archetype, risk tier, estimated hours, and the client’s industry.
- The fit gate runs. We publish the eligible roster for that project. Eligibility is the matching rule plus being free of any other engagement. If you are not eligible, you cannot bid and you lose nothing.
- One seat is appointed, not auctioned. Faculty names a builder rated 3 in the primary archetype and confirms the two-deep backup.
- The other seat is auctioned. Sealed bids from the eligible roster. One bid each, 72-hour window, no revisions.
- Second price wins. The highest bid takes the seat and pays the second-highest bid plus one point. Everyone else pays nothing and keeps their balance.
- Ties break toward the builder with fewer completed engagements, then by coin flip.
- Results are published. Every bid and every clearing price, after the fact. You should be able to see what projects have historically cost.
Bid your actual level of interest. Second-price is not a gimmick. Because the winner pays the runner-up’s bid rather than their own, there is no advantage to shading your bid downward or trying to guess what everyone else will do. The strategy that wins is stating what the project is genuinely worth to you. If you find yourself doing game theory, you are working harder than the rules require.
Guardrails
Fit is never overridden by points. A large bid from an ineligible builder is not a bid. The client is paying for the right team, and enthusiasm does not substitute for the skill they are buying.
Nobody sits out a semester. If you have zero engagements by the midpoint of the term, you get a directed placement into the next opening, outside the auction. The market allocates enthusiasm. It does not decide who gets to participate.
Some projects are closed to the auction. Engagements involving regulated data, records about minors, or systems that influence decisions about people are staffed by appointment. There are not many of these, and you will know which ones they are because the brief says so.
Faculty veto exists. It gets used rarely and it gets explained when it does.
Discovery Teams
A discovery-only engagement is not a full pod. It is an engagement lead plus one builder, for as long as the client agrees, and its job is to find out what the real build is and how long it will take.
If the build phase then signs, the builder who ran discovery gets first claim on the appointed seat without bidding. The context they built is worth more than the scheduling convenience of reopening it.
Discovery is the fastest route onto a project you care about. It costs no points.
What This Means for You
- Your archetype ratings decide your eligibility, so raise the low ones. The fastest way to become eligible for more projects is to ship something in an archetype where you are a 1.
- Being a 3 in something matters more than being a 2 in everything. Rule 1 asks for a specialist. Rule 4 means we need two of them.
- Spend points on the project you actually want. Hoarding to 300 and never bidding is the one strategy that is strictly dominated.
- Take a discovery slot if one is open. It is a free option on the build seat.