Phase 1 — The Moment
Handshake AI posted a Staff Product Designer for its tasking platform on September 23, three days ago. It is one piece of a design org being assembled in public: CPO Vee Sahgal announced a VP of Product search on September 15, and two earlier design requisitions have since left the active Ashby board — a Founding Staff role for Enterprise, off the feed since July 30, and an operator-independence Staff seat posted at $190K–$240K. Filled, withdrawn, or folded into this one: public signal doesn't say.
Real mandate: own design for the internal platform that determines how Handshake's AI models get trained, evaluated, and improved. The posting lists dependencies across Quality, Coding, RLE, Fellow Experience, research, operations, and data — those are Handshake's internal team names, not functions you'd recognize from outside. This is model-quality infrastructure. The people using it are internal operators, not end users.
The comp band moved. The July requisition posted $190K–$240K. This one posts $225K–$250K. Both numbers are sourced from the postings; the reason for the gap is inference. Two readings fit: the scope expanded, or the first search failed at the lower band and this posting reflects what the committee learned it actually needed. Either way, the seat is worth more to them now than it was in July.
Sahgal's VP Product post says that executive will own strategy "across modalities, our operator and fellow platforms, and customer-facing surfaces." No VP Design or Head of Design appears anywhere in Handshake's public org. Design authority most likely sits inside product leadership. That is the gate question for this pursuit, and it is why the outreach below is written the way it is.
Issue #11 covered a different Handshake seat with a different evidence center — agent authority, Trust essay leading. This requisition is an operator-workflow and platform-design problem. Different seat, different mapping.
Rubric:
| Company | Score | Note |
|---|---|---|
| AI centrality | +2 | AI is a primary strategic bet; Handshake itself is an employment platform |
| Stage / equity | 2 | Late-stage pre-IPO (Series F, $3.5B valuation Jan 2022); equity window narrowing |
| Design influence ceiling | 1–2 | No visible design executive; reporting line unconfirmed; VP Product search suggests design sits inside product |
| Trajectory | 2 | Active org buildout, but roles appearing and disappearing signals restructuring |
| Company total | 7–8 |
| Role | Score | Note |
|---|---|---|
| Comp potential | 2 | $225K–$250K base; equity terms unknown |
| Scope expandability | 2 | Broad platform mandate; ceiling depends entirely on where design reports |
| Craft depth | 3 | Staff IC, hands-on, end-to-end ownership |
| AI exposure quality | 3 | Designing the infrastructure that shapes model training and quality |
| Portfolio value | 2 | Operator-facing AI platform; differentiated but internal |
| Role total | 12 |
Role clears 12. Company falls short of 9 on one unresolved variable: the design influence ceiling. The outreach exists to test that gate. Reverts to Watch if Sahgal's reply confirms design reports through product with no independent authority.
Phase 2 — Portfolio Mapping
Intelligence layer check: the tasking platform surfaces model-quality data to internal operators, which is enough to trigger the check. It doesn't survive it. The design surface here is operator workflow and exception handling, not the moment between what a model surfaces and what a person decides. Thermo Fisher leads.
Lead with Thermo Fisher mySupply. The tasking platform spans six functional areas. mySupply spanned six pharma partners across nine sites. Same shape of problem: teams with competing operational priorities on one platform, where the design job is surfacing the exceptions that matter — three orders out of 197 — rather than displaying everything at equal weight. mySupply was the first system where all six partners saw shared performance data. Handshake needs that cross-team visibility for model quality, task completion, and operator throughput. The $20M margin opportunity, 100% partner adoption, and 83% IRR are the evidence those decisions converted.
The Thermo Fisher case page doesn't name BCG Digital Ventures. The portfolio homepage does. Put the attribution on the resume explicitly.
Currency: TinyFish. Most recently as Head of Product at TinyFish, Juno shipped an enterprise web-agent platform from 0 to 1 in three months and worked daily through agent traces, auditability, attribution, and governance in production deployments. That is practitioner depth a portfolio-only candidate can't claim, and it applies directly to a mandate that shapes model behavior through data operations and quality workflows.
Support: Brand Pulse. Live and inspectable. Parallel agents pulling Reddit and X, classifying every mention by sentiment and theme, building a live brand score. A reviewer can open it in a browser and see the system working. That is what makes it useful here: it corroborates the TinyFish claim with something a stranger can check.
Hardest objection: "The tasking platform is ML infrastructure — task design, quality evaluation, RLHF workflows, data operations. Your background is enterprise product design." It's hard because the posting's team dependencies read like an engineering org, and the buyer may be screening for someone who already speaks that vocabulary. Thermo Fisher is the answer: a supply chain platform whose entire design problem was making batch tracking, exception handling, and compliance events actionable for operators who weren't supply chain engineers. The translation problem is the same. The TinyFish year covers the difference in substrate.
Phase 3 — The Outreach Package
Why Sahgal: no design executive is visible in Handshake's public org and no hiring manager is named on the posting. Sahgal is the nearest verified hiring authority for this seat.
First contact message — to Vee Sahgal:
Vee — your post about the VP Product search described that role owning strategy across modalities, the operator and fellow platforms, and customer-facing surfaces. I'm trying to understand where the September 23 Staff Product Designer for the tasking platform sits relative to that incoming VP, and whether it's a newly scoped seat or an evolution of an earlier one.
I ask because the mandate — design across Quality, Coding, RLE, Fellow Experience, and operations — is close to work I led as Product Design Director at BCG Digital Ventures. On Thermo Fisher's mySupply platform I designed exception-first workflows across six pharma partners and nine sites, where the problem was surfacing the three orders that mattered out of 197 for teams with competing priorities. It reached 100% partner adoption and $20M+ in margin.
Most recently I was Head of Product at TinyFish, where I shipped an enterprise AI agent platform from zero to production and worked daily through agent traces, auditability, and governance in deployment. The AI-infrastructure half of this mandate isn't abstract to me.
Would 20 minutes make sense to test whether the scope and reporting line match what I'd bring?
(188 words)
Resume framing note:
Open the summary with the Thermo Fisher mandate — multi-stakeholder platform design, exception-first workflows, operator research inside complex operational environments — and surface the $20M margin and 100% adoption figures before anything else. Carry BCG Digital Ventures in the summary line and beside every BCG DV project; the case page doesn't, so the resume has to. TinyFish goes second, stated flat: "Head of Product, TinyFish — shipped enterprise web-agent platform 0→1, 3 months." Alibaba goes third and stays subordinated, since B2B platform scale is adjacent but this platform is internal infrastructure. Cut Equinox+ and Allē; consumer loyalty puts distance between her and an operator-workflow seat. Keep "design leadership" out of the summary entirely. This is a Staff IC role, and the buyer is screening for hands-on platform judgment.
Cover letter hook:
The tasking platform's mandate — design across Quality, Coding, RLE, Fellow Experience, and operations — is the same multi-stakeholder, exception-first problem I solved on Thermo Fisher's mySupply, where six pharma partners with competing priorities shared one platform for the first time. What's different is the substrate: model quality instead of batch compliance, which is the gap I spent the last year closing by shipping AI agents in production at TinyFish.
Phase 4 — Window Summary
| Action | Deadline | What degrades without it |
|---|---|---|
| Send Sahgal the note testing reporting line and mandate scope | Monday September 28 | The posting turns five days old. First-mover advantage on a fresh req erodes daily; every additional day raises the odds a recruiter screen has already narrowed the pool. |
| Build no bespoke artifact until Sahgal confirms the mandate. Reporting line, relationship to the July requisition, and design authority boundaries are all unresolved. | Hold until reply, or 10 business days (October 10) | A custom artifact for a seat that may report through product with no design influence is preparation spent on a role that scores Watch. If no reply by October 10, reassess whether the channel is still open. |
| If Sahgal confirms: one page mapping Thermo Fisher exception-first design onto ML quality workflows | Within 48 hours of confirmation | Without it the portfolio reads as enterprise design with recent AI exposure. With it, the portfolio reads as operator-platform design whose last year was AI infrastructure. |
October 10, 2026.
- Sahgal's "signal broke" thesis: His March post introducing AI Showcase argued that listing AI skills on a résumé no longer demonstrates what a candidate can build — useful framing if the conversation moves toward how Handshake evaluates proof of work internally.
- Enterprise founding seat context: The archived Founding Staff Product Designer, Handshake AI Enterprise posting described a different design surface — external customer deployments, enterprise pilots, sales collaboration, and building a design function from zero — which maps more closely to Alibaba than to Thermo Fisher if that mandate resurfaces.
- Handshake's "refounding" strategy: Their public essay on becoming a career network for the AI economy declares a shift from static applications toward verified task-based experience and conversational agents, which is the strategic context behind the tasking platform but should be attributed as company intent rather than shipped product.
- Webflow Brief as outreach precedent: The Agentic CMS Brief in issue #12 established the "TinyFish in conversation only" wording and confirmed the past-tense positioning formula — worth re-reading before sending the Sahgal note to keep the TinyFish framing consistent across concurrent outreach.
