A practicum proposal · 2026/27 For Master's Faculty
The She & HER Institute

Your master's students belong on this problem.

We're building a credit-bearing one-semester practicum for master's students in MBA, behavioral economics, data science, and mathematics — placing them on live operational, research, and engineering work inside a venture-backed consumer platform serving sapphic women (women who love women).

I.
The Opportunity

A consumer platform serving a population the research field has under-instrumented.

She & HER is a sapphic dating platform — for women who love women — built on a deliberate product thesis: pedagogy, not pairing. The product delays romantic matching and instead routes users through structured community, self, and friendship surfaces first.

The technical and behavioral problems that follow from that thesis are interesting in the academic sense, not just the product sense. Matching systems for sparse identity-defined populations. Behavioral interventions designed to delay rather than maximize engagement. Fairness metrics where the protected attribute is the entire product. Retention modeling for cohorts whose first-touch interaction is not romantic.

None of this is well-covered in the recommender-systems or behavioral-economics literature. Mainstream dating-app research has been built around heterosexual-default products and pairing-maximizing objective functions. The sapphic population — and the broader question of what a non-extractive consumer dating product can look like — sits outside that frame.

A dating product that teaches its users to date well, instead of optimizing for engagement at any cost — that is a research frontier, and the field hasn't named it yet.

Beta launch is September 2026, with a coalition distribution motion that already runs through MBA programs, HBCUs, and identity-coded community partners. The product is real. The data will be real by Fall semester. Your students would be working on a live system serving real users — not a case study.

0
academic tracks for v1
0
week practicum, 3 credits
0
student pilot cohort, fall 2026 / spring 2027
II.
The Program

A one-semester practicum with a clean academic shape.

The Institute is designed to fit the practicum machinery your program already runs. One semester. Three academic credits. Twelve to fourteen weeks. A faculty advisor of your designation owns academic admission, grading authority, and university-side compliance. She & HER provides the industry mentor, the real problem, the dataset access, and a structured weekly cadence the faculty advisor can rely on without micromanaging.

Two students per partner institution per cohort in the pilot year. That number is deliberate. It protects the supervision time our small team — a founder/CEO with a fifteen-year ML and data-science background, a CTO from Google and an enterprise GenAI engineering role, and a product manager — can commit without breaking. We can grow to four per school once the platform has launched and team capacity is larger.

Grading is the faculty advisor's call. We submit a written performance evaluation that maps to the rubric in Week 13. The mapping protects against the failure mode where a student does excellent corporate work but produces no academic artifact — or vice versa.

Faculty advisors are offered co-authorship on resulting research publications and named institutional affiliation on the program. Once the publication portfolio is established, faculty advisors become natural co-PIs on grant applications — which is where the Institute pays for itself in Year Two.

Application: standard practicum channel + one-page fit memo to us. Selection: faculty shortlists; we conduct a 30-minute fit interview with finalists. Onboarding: half-day orientation in Week 1, NDAs signed, mentor office hour before week's end.

III.
Four Tracks

The work your students would actually do.

Each track is built around a live business or research problem that is unblocked, instrumented, and ready for student work by Fall 2026. Click any track to expand.

A.
MBA Strategy & Go-to-Market
Mentor: Desirée Mayon, CEO & founder

Live problem

Each student takes ownership of one segment of the She & HER coalition GTM motion (MBA distribution at top business schools, HBCU coalition, Pride-organization partner network, or brand partnership pipeline) and builds the strategy, partnership economics, and ninety-day execution plan for that segment ahead of the September 2026 beta.

Learning objectives

  • Structure coalition partnerships with non-traditional partners where standard B2B SaaS playbooks don't apply.
  • Model partnership economics where consideration is rarely cash — audience access, co-branded content, equity carve-outs, in-kind contributions.
  • Translate founder-led pitching into repeatable team processes: the bridge between Stage 0 founder magic and Stage 1 sales motion.

Deliverables

  • Week 6: Segment GTM playbook (5–10 page document) for sign-off by industry mentor and faculty advisor.
  • Week 10: Partnership economic model with quantified cost and return projections per partnership type.
  • Week 14: Executive readout to founder team. Top performers present at the next quarterly board update.
B.
Behavioral Economics
Mentor: Desirée Mayon + Sevanne (Product)

Live problem

She & HER's product thesis — pedagogy not pairing — has to be operationalized into measurable behavioral interventions and tested against real cohorts. Students design reciprocity nudges, structured delay mechanics, and desire-articulation prompts; instrument them; ship them behind feature flags; and analyze the results.

Learning objectives

  • Design behavioral interventions for a consumer product where the metric of interest is not engagement-maximizing.
  • Run A/B tests in a dating-app context where standard confounds — network effects, identity-driven preference shifts, small N for under-represented groups — make naive analysis misleading.
  • Apply behavioral-economics principles to pro-social product design rather than dark-pattern outcomes.

Deliverables

  • Week 4: Intervention design document for two candidate interventions, mentor-reviewed.
  • Week 9: Live A/B results from at least one intervention with effect-size estimates and confidence intervals appropriate to the cohort N.
  • Week 14: White paper draft (8–15 pages) suitable for CHI, CSCW, or a NeurIPS workshop submission. Joint authorship.
C.
Data Science & Machine Learning
Mentor: Mike Castner, CTO (ex-Google, Pelidum) + Desirée

Live problem

The matching system uses an explore-vs-exploit foundation with streaming re-rank in batches of ten. Three open extension areas, each both academically interesting and operationally load-bearing: fairness metrics for matching in sparse identity-defined populations, cold-start solutions for users outside dominant preference clusters, and retention modeling for cohorts whose first interaction is intentionally non-romantic.

Learning objectives

  • Build, deploy, and evaluate components of a production matching system serving real users.
  • Apply fairness metrics from the recommender-systems literature in a context where the protected attribute is the entire product surface.
  • Model retention as a function of product structure, not engagement metrics — particularly for users whose progression is non-linear.
  • Operate in a production engineering environment: Bazel monorepo, gRPC services, Cloud SQL Postgres, code review, CI/CD discipline.

Deliverables

  • Week 5: Technical design doc reviewed by CTO.
  • Week 10: Working implementation committed to private repository with offline evaluation results.
  • Week 14: Production deployment behind feature flag where appropriate, with live results memo. Joint authorship on resulting paper.
D.
Mathematics
Mentor: Desirée Mayon (MS Bioinformatics, 15+ yrs applied probability)

Live problem

The product thesis — pedagogy not pairing — is qualitative. Track B turns it into behavioral interventions; Track C ships and measures them. The bridge in the middle is the formal mathematical model: given these assumptions about user utility, what are the testable predictions, what are the boundary conditions, and what would falsify the thesis? The Math track aligns the qualitative with the quantitative. Students choose from four problem options in Week 2 — multi-tier matching theory, mathematical formalization of reciprocity nudges, measurement theory for sparse identity-defined populations, or information-theoretic bounds on profile-to-match utility — or propose an alternative.

Learning objectives

  • Translate qualitative product and behavioral hypotheses into formal models with stated assumptions and testable predictions.
  • Connect mathematical theory to live product instrumentation. The model isn't theory-for-its-own-sake; Tracks B and C are downstream consumers of the predictions it generates.
  • Apply measure-theoretic, game-theoretic, or information-theoretic tools to applied problems in consumer products where the data-generating process is non-standard.
  • Frame and write findings to a standard usable for mathematical psychology, operations research, decision science, or theoretical economics venues.

Deliverables

  • Week 2: Problem-selection memo with stated assumptions and predictions the model will generate.
  • Week 7: Model draft — formal statement, key results, empirical predictions. Industry-mentor review.
  • Week 14: Working paper (10–20 pages) suitable for peer-reviewed submission (Theoretical Economics, Games and Economic Behavior, Mathematical Psychology, Operations Research, or a NeurIPS/ICML theory workshop). Joint authorship.

Cross-track interaction

This track is deliberately load-bearing for Tracks B and C. The Math student's mid-semester model becomes a working hypothesis for the BehavEcon student to test or a theoretical bound the Data-ML student designs against. Weekly Friday demos are joint across all four tracks; the cross-pollination is part of the curriculum, not a side effect.

IV.
The Semester

A fourteen-week ledger. The four tracks hand off across it.

The tracks are not parallel silos. The Math model becomes a working hypothesis for BehavEcon to test; the BehavEcon result becomes a theoretical bound the Data-ML student designs against; the MBA student distributes the resulting product. Tap any week to see what's happening across all four tracks.

i.
Onboarding week. All tracks
Orientation Half-day session with the founder team. Program intent, expectations, weekly cadence.
Provisioning NDAs signed. Notion workspace access. Track C: read-only repo access provisioned.
Reading list Three to five track-specific items assigned. Brevity is the rule — these are not survey papers, they are the anchors.
First office hour With named industry mentor before end of week.
+
ii.
Math problem-selection memo. Math
Math Student selects one of four pre-scoped problems (multi-tier matching theory, reciprocity-nudge formalization, sparse-population measurement theory, information-theoretic match-utility bounds) or proposes an alternative. Memo states assumptions and the testable predictions the model will generate.
Other tracks Scoping continues. BehavEcon begins intervention design. Data-ML begins technical design.
+
iv.
BehavEcon intervention design document. BehavEcon
BehavEcon Two candidate interventions designed and documented. Reviewed by Desirée and Sevanne. Approved interventions move into instrumentation.
Math hand-off Math student's Week 2 memo informs which interventions are tractable to formalize. The two students consult.
+
v.
Data-ML technical design doc. Data-ML
Data-ML Technical design for the chosen extension area (fairness metrics, cold-start, or non-linear retention model). Reviewed by Mike Castner.
Math hand-off If the chosen Math problem is information-theoretic bounds or multi-tier matching theory, the Math student's framing constrains the Data-ML student's design space.
+
vi.
MBA segment GTM playbook. MBA
MBA Five-to-ten page GTM playbook for one coalition segment (MBA distribution, HBCU coalition, Pride-organization network, or brand partnership pipeline). Sign-off by industry mentor and faculty.
Other tracks All four tracks present a midpoint demo at the Friday all-hands.
+
vii.
Math model draft. Math
Math Formal statement of the model. Key results. Empirical predictions. Reviewed by Desirée.
Cross-track The model's predictions are shared with BehavEcon (to inform what the live test should measure) and Data-ML (to inform what the production system should satisfy).
+
ix.
BehavEcon live A/B test results. BehavEcon
BehavEcon At least one intervention has run. Effect-size estimates and confidence intervals reported, with attention to small-N corrections appropriate to the cohort.
Math hand-off Results either confirm, qualify, or falsify the Math student's predictions. The two students co-write a short integration memo.
+
x.
MBA economic model. Data-ML implementation complete. MBA Data-ML
MBA Partnership economic model — quantified projections of cost (founder time, marketing spend, equity, in-kind) and return (acquisition, retention, brand value) per partnership type.
Data-ML Working implementation committed to private repo with offline evaluation results.
Math Begin paper write-up.
BehavEcon Begin white paper.
+
xiii.
Performance evaluation to faculty. All tracks
She & HER Written performance evaluation submitted for each student, mapping to the faculty advisor's rubric. Captures both corporate work product and academic deliverable quality so the rubric handles the edge cases cleanly.
Faculty Final grade is the faculty advisor's call. The evaluation is an input, not a substitute.
+
xiv.
Final readouts. Working papers. Production deploy. All tracks
MBA Executive readout to founder team. Top performers present at next quarterly board update.
BehavEcon White paper draft (8–15 pages) — conference-submission-ready.
Data-ML Production deployment behind feature flag with live results memo.
Math Working paper (10–20 pages) — peer-review-submission-ready.
+
V.
The Swap

What we offer your students. What we ask of your program.

Year One is a clean credit swap. No tuition flows to She & HER. No stipend flows from She & HER. Year Two opens the grant flywheel.

From She & HER

What students & programs receive

  • A live product problem, not a case study. Real users, real data, real ship dates.
  • An industry mentor per track with FAANG, AI engineering, or founder-level depth in the relevant discipline.
  • A portfolio artifact — shipped code, a deployed intervention, a real GTM playbook — that is verifiably the student's own work.
  • A written performance evaluation on the faculty advisor's rubric, delivered Week 13.
  • A reference letter from the CEO and founder for strong performers, available for PhD applications, fellowships, or hiring processes.
  • Co-authorship on resulting research publications for students whose work qualifies.
  • Named institutional affiliation for the partner university on resulting publications and the program itself.
From the program

What we'd need from your university

  • A faculty advisor of your designation — practicum coordinator, tenure-track faculty, or clinical/professor-of-practice. One named person per cohort.
  • Course-catalog placement for the practicum in your normal channel, or independent-study authorization for the pilot year.
  • Academic admission control — the faculty advisor shortlists; we conduct a 30-minute fit interview with finalists.
  • Standard university IP and confidentiality framework for industry-anchored coursework. The shape is conventional — work product belongs to She & HER; students retain academic-use rights; publications go through a 30-day review with co-authorship for material contributors. Your office of general counsel will recognize it.
  • For Track C students: CPT authorization for any international students participating in unpaid practica.
Year Two and beyond

The Year One cohort produces publication-worthy research and establishes formal university research-site relationships with named faculty PIs. Beginning Year Two, we apply jointly for NSF (Human-Centered Computing, Information Integration and Informatics), NIH (sexual and gender minority health research portfolios), and foundation grants (Ford, Mellon, Mozilla, MacArthur, and discretionary funds from program-affiliated research centers). She & HER as PI or co-PI; partner university as research site, or vice versa. By Year Three the program is grant-funded and the supervision time is paid for.

VI.
Who's Leading

Built by a founder with the credentials this program asks of itself.

Desirée Yvonne Mayon
CEO & Founder, She & HER
  • 15+ years in machine learning and data science
  • Google · Software engineer and ML research
  • Microsoft / Xbox · Data science
  • University of Cambridge · Research affiliate
  • Etsy, Nordstrom · ML applied to consumer products
  • MS Bioinformatics · Texas A&M University
  • Black sapphic founder leading the company

The Institute is being proposed by a founder who has been on both sides of the industry-academic table. Fifteen-plus years in machine learning and data science across Google, Microsoft, Etsy, Nordstrom, and Cambridge means we know what student work looks like when it's genuinely useful to a company, and what it looks like when it's busy-work that wastes a master's-level semester.

Mike Castner, our CTO, built the MPAC GenAI engine at Pelidum and spent nine years at Google across Safe Browsing, Fuchsia, Spanner, and the Android Privacy Sandbox. He earned fifty-eight peer-recognition awards in that span. He's the named technical mentor for Track C.

The product thesis we're operationalizing — that a sapphic dating platform should teach reciprocity, needs articulation, and desire-reading before it pairs anyone — was not invented in a vacuum. It came from a community that has been organizing its own dating infrastructure for generations. We treat the field history with the seriousness it deserves; we expect students who join the practicum to do the same.

A note on the team's bandwidth.

This is a small team carrying a venture-scale product into a beta launch in September 2026. We have done the supervision math: a two-student-per-school cohort across four partner institutions costs the founder team approximately twelve hours per week of mentorship and review during the semester. We will not over-commit. If a school's calendar requires Spring 2027 to do this cleanly, that is the right answer.

VII.
Anticipated Questions

The questions we'd ask if we were on your side of this conversation.

What if we don't have a faculty member with sapphic dating research expertise?

The faculty advisor does not need topical expertise in sapphic dating. The expertise required is in the academic discipline — operations research, behavioral economics, machine learning, mathematics — and in the practicum format itself: how to run industry-anchored coursework, how to evaluate student work that has both corporate and academic components, how to navigate university IP frameworks.

The substantive research direction comes from the industry mentor. The faculty advisor anchors the academic shape and the grading. A faculty member who has run a master's-level practicum or capstone with any industry partner has done this work before.

How do we handle IP when a student wants to use practicum work for their thesis?

Standard for industry-anchored coursework. Work product belongs to She & HER, Inc. The student retains a perpetual, royalty-free right to use their own contributions for academic purposes — thesis chapters, course requirements, portfolio, conference presentations.

For publications, manuscripts go through a 30-day review with She & HER. Co-authorship is offered to She & HER personnel and faculty advisors who contributed materially. We are happy to share the full IP and confidentiality framework with your office of general counsel before the program is formalized.

What if Fall 2026 doesn't fit our course catalog?

Spring 2027 is the right answer for any program where Fall 2026 would require rushed approvals. We would rather start later and start cleanly. The curriculum is identical; only the calendar shifts.

For programs that can accommodate Fall 2026, an independent-study placement under your faculty advisor's discretion is often the cleanest path while a formal practicum slot is still being added to the catalog for subsequent semesters.

How are international students on F-1 visas handled?

The practicum is unpaid in Year One. F-1 students participating typically need CPT (Curricular Practical Training) authorization from your designated school official, which most master's programs handle as a routine matter for practicum coursework. Your faculty advisor and DSO are the right pair to confirm specifics at intake.

If CPT authorization becomes a blocker, we have flexibility on the structure — an OPT-aligned timeline or a research-credit alternative can usually accommodate.

What does the supervision actually look like from the student's side?

One scheduled office hour per week with the named industry mentor for the student's track. Asynchronous review of work product (memos, code, designs) within 48 hours of submission. A Friday all-tracks demo where the eight students present briefly to the founder team and to each other.

Onboarding week is half-day in-person if geography allows, otherwise structured video. Final week includes a written readout and an executive presentation to She & HER's founder team — for top performers in the MBA track, that readout extends to the company's quarterly board update.

How is sensitive user data handled given the population the product serves?

This is the question we want faculty to ask. Students access pseudonymized production data only. No personally identifying information leaves the production environment. The 30-day data retention policy that governs She & HER's main product applies to all practicum work.

Per-student data-access scope is approved by the founder and CTO before semester start. The data framework will not be the bottleneck — it is one of the most considered parts of the product, given who we serve.

What's in it for our faculty advisor specifically?

Three things, in increasing order of commitment.

Named institutional affiliation on the program itself — guaranteed. The faculty advisor's school is named on the masthead of any program-level material and on every publication that emerges from their cohort's work.

Co-authorship on resulting research publications — offered for material contribution. Faculty advisors who actively supervise, redirect, or shape the work are normally co-authors on the resulting paper, and this practicum follows that convention. We do not bury authorship offers; we name them upfront.

Co-investigator status on Year Two grant applications — offered for material contribution to the grant itself, and where the faculty advisor's discipline and institutional fit are right for the funder. The grant flywheel is the medium-term economic answer for both sides; Year One is the investment, Years Two and Three pay for it.

Beyond the formal incentives: the master's program gains an industry-anchored practicum with a high-credibility partner, which has historically been a draw for prospective students.

What if a student drops out mid-semester?

Standard academic-side handling — the student's grade is determined by the faculty advisor per the program's existing policies. On the She & HER side, work product produced through the date of departure stays with She & HER per the IP agreement; the student retains academic-use rights on their own contributions.

For the cohort, the practical effect is small. The remaining students continue; the cross-track hand-offs adjust around the gap. We do not backfill mid-semester.

A 30-Minute Call

Let's see if the shape fits your program.

The next step is a thirty-minute conversation to map the program against the shape of your master's curriculum, identify the faculty members who would be the right fit to anchor it, and settle the timing — Fall 2026 if your catalog allows, Spring 2027 if cleaner.