# Clay — Strategic Finance / FP&A | note to Karan Parekh (hiring manager) — 2026-06-09

Role: Strategic Finance - FP&A (Ashby `claylabs`, jid 11523b92) · Clay · NYC, onsite · $150–200K + equity
Apply: clay.link/8633jfp
Resume: `final resumes/2026-06-09/Clay/Strategic Finance FP&A/Darsh Shah - Clay Strategic Finance FP&A Resume - 2026-06-09.pdf`

Karan's post (the hook): he doesn't want a traditional FP&A "number-puller" — wants "learn → translate → drive the business," owner's mindset. His example: Clay was losing money charging a FIXED price for VARIABLE AI usage cost → modeled and shipped variable pricing cross-functionally. Darsh did exactly this at Epic.

## Note (LinkedIn message / connection note) — paste

Hi Karan — your FP&A post resonated, especially "we don't want another number-puller."

The example you gave — Clay losing money on a fixed price for variable AI usage cost, then modeling and shipping variable pricing — is almost exactly what I've owned at Epic. I built the consumption-based pricing model (per-call, cache-hit, usage) for our AI creator tools: repriced to the underlying cost curve, protected ~50% gross margin, and took ~$25M of annualized margin risk off the table.

Beyond the model, I run our planning cadence and built finance's first agentic tooling on Claude Code so partners self-serve their own analysis — learn → translate → drive is already how I operate. Kellogg MBA, ~6 years in strategic finance. Worth 15 minutes?

— Darsh

## Shorter variant (if you want it tighter)

Hi Karan — your FP&A post resonated. Your example (Clay charging a fixed price for variable AI usage cost → modeling + shipping variable pricing) is almost exactly what I owned at Epic: I built the consumption-based pricing model for our AI creator tools, protected ~50% gross margin, and removed ~$25M of annualized margin risk. I also run the planning cadence and built finance's first agentic tooling on Claude Code. Kellogg MBA, ~6 yrs strategic finance — learn → translate → drive is how I already work. Worth 15 minutes? — Darsh

## COLD message (recommended — first contact, no intro)

Hi Karan — saw your post on the Clay FP&A role, and the part about not wanting a number-puller (plus your example of pricing variable AI cost) is the exact problem I work on day to day.

At Epic I own the consumption-pricing and unit-economics model for our AI creator tools — pricing per-call/cache-hit/usage so variable inference cost doesn't quietly eat margin (~$25M of annualized risk off the table, ~50% gross held). I also run the planning cadence and built our finance team's first agentic tooling on Claude Code, so partners self-serve the analysis instead of waiting on reports.

Kellogg MBA, ~6 years in strategic finance. If the role's still open, I'd value 15 minutes.

— Darsh

### Connection-request note (≤300 chars, if you connect first)

Hi Karan — your Clay FP&A post resonated, esp. "not another number-puller." At Epic I own consumption-pricing + unit economics for our AI creator tools (pricing variable inference cost so it doesn't eat margin, ~$25M risk removed) and built finance's first agentic tooling. Kellogg MBA — would love to connect.

_Truthfulness note: framed as "built the consumption-pricing model" (true), NOT "fixed an existing flat-fee mispricing" — only use the stronger fix-it framing if that's actually what happened._