Maple builds conversational AI agents that handle phone calls for restaurants and hospitality venues—handling orders, reservations, and payments synced directly to existing POS systems. The stack reveals a compute-heavy ML operation: PyTorch, JAX, CUDA, Triton, and GPU/TPU clusters power the voice models, while Twilio handles PSTN integration and WebSockets manage real-time agent interactions. Founded in 2023 and now scaling toward 40k customers, the company is engineering-dominant (8 of 11 hires in recent months) and actively tackling latency and noise-robustness challenges typical of edge voice deployments.
Maple provides voice AI agents designed for phone-first businesses—primarily restaurants, drive-thrus, and kiosks. The agents answer calls, process orders, book reservations, and handle payments, with data synced back to the merchant's existing systems (via Salesforce and Twilio integration). The company is pre-Series A scale (11–50 employees, founded 2023) and headquartered in New York. Current product focus spans agent interaction design, real-time conversation quality, and rapid customer onboarding; longer-term expansion targets drive-thru and kiosk variants. Active hiring is weighted toward engineering and support, signaling both infrastructure scaling and customer deployment complexity.
PyTorch, JAX, CUDA, Triton, and GPU/TPU clusters for model inference; TypeScript, React, Node.js, PostgreSQL for backend; Twilio for PSTN, Salesforce for CRM, Slack for internal tooling.
United States and South Africa are both active hiring regions.
Other companies in the same industry, closest in size