Vol. VI, No. 2026 ◆ Greater Seattle Area
Est. 2019
Price: One Interview Call
Heni Bhungalia
"Build for the second visit, not the first." · Build something worth coming back to.
Seattle, WA. Heni Bhungalia shipped four production AI features on DreamCRM, drove onboarding from 17% to 50%, and achieved 57–60% month-1 retention, more than double the B2B CRM industry average.
She led a cross-functional team of 5 engineers, 4 interns, and 1 marketer across DreamCRM, ZeroGPU, and Unirac Solar concurrently. ZeroGPU secured pre-seed and seed funding on the portal strategy she shaped. The Unirac Solar portal redesign she led collapsed a 22-step workflow to 8 inputs with live editable output, cutting interactions by 55 percent and contributing to $105M+ in YTD sales.
Built from zero: Gmail Sync, LinkedIn Sync, AI-suggested actions, AI email drafting. Month-1 retention: 57–60% vs a 20–30% industry average. Onboarding: 17% to 50%. Avg. session length: 7m 41s.
Cleared Google CASA Tier 2 OAuth, a B2B sales eligibility gate most early-stage teams skip entirely.
Defined Developer and Edge Operator portal strategy for ZeroGPU, an edge AI infrastructure platform. Validated API architecture via Postman. Authored product roadmap that directly contributed to ZeroGPU securing pre-seed and seed funding.
Redesign of the installer and distributor portal. Reduced a 22-step BOM generation workflow into a single 8-input dashboard with live editable output, cutting installer interactions by 55 percent and contributing to $105M+ in YTD sales. Two brainstorm prototypes and a design review aligned stakeholders on the new-project flow, landing on a hybrid setup with reusable templates so repeat installers skip configuration. Led cross-functional team across multiple time zones by aligning stakeholders and leveraging existing product data.
Led a 0-to-1 financial intelligence product end-to-end, translating UBPR and FRED datasets into metrics for community and mid-size banks. Ran weekly discovery with 3 banking design partners to ship a dashboard and AI chatbot for natural-language queries grounded in the bank's own data. Now in active review with bank executives.
Users describe an appliance problem in plain language; the system asks targeted follow-up questions and returns a repair vs. replace quote grounded in domain-specific APIs. More accurate than generic AI tools, with full diagnosis transparency. Users can request a verified technician directly. Delivered beyond MVP; technician portal in progress.
Eliminated the manual pre-call research grind for property BD agents. The system aggregates every public listing for a property, analyzes what is working and what is not, and surfaces actionable insights before the call so agents skip basic questions and have meaningful conversations from minute one. MVP shipped; expanding.
Led end-to-end development of CAMD's public research platform through a US–Italy collaboration. WCAG-compliant, cleared accessibility and brand audits. Led the DRS 2024 paper alongside senior faculty co-authors: 1,241+ downloads in year one.
Built customer-facing features for Stop & Shop and Food Lion (300K+ combined daily visitors), including Weekly Ads, Savings History, and search/discovery components that supported a 30% conversion lift in A/B tests. All within a 6-month co-op.
Build something worth coming back to.