You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
SafeRide builds open-source AI safety infrastructure for physical-world transportation. We run real-time fatigue/distraction detection on-device via MediaPipe CV inference, capture vehicle telemetry with Pico W + IMU kits, and anchor tamper-evident incident logs on-chain for trustless verification.
Goals: (1) Deploy 10-driver live pilot in Guwahati to validate tamper-evident safety logging under real-world conditions, (2) Release open-source DePIN storage SDK for content-addressed safety archives, (3) Publish security benchmarks and threat model documentation for edge AI safety systems.
How funding will be used:
$10K minimum: Hardware kits for 10 drivers ($3K), driver stipends ($2K), on-chain transaction costs ($1K), SDK documentation and testing ($2K), contingency ($2K).
$25K goal: All above plus DePIN storage SDK development ($5K), insurer/fleet validation testing ($3K), performance benchmarks and security audit ($3K), community outreach ($2K), extended pilot ($2K).
Team and track record:
Two technical founders. I'm ex-Hugging Face ML researcher, current Research Fellow at University of Nottingham, LFX'26 mentee at Linux Foundation. Built SafeRide's full stack from scratch: Go backend, MediaPipe CV pipeline, Solana logging, SvelteKit dashboard. Co-founder handles backend and storage infrastructure. Built AQUAZAP (Android pentesting framework, 100+ attack vectors). Pilot design complete with 10 drivers identified in Guwahati.
Most likely causes of failure:
(1) Hardware supply chain delays in Northeast India could push pilot timeline, (2) Driver retention during 1-month pilot if stipends are insufficient, (3) On-chain logging costs spike if transaction volume exceeds estimates. Mitigation: hardware pre-ordering, driver agreements signed upfront, Solana's low fees provide buffer.
If project fails: SafeRide stays a prototype on GitHub. Open-source code remains available but unvalidated. No live safety data to demonstrate tamper-evident verification works in adversarial conditions. DePIN SDK and security benchmarks delayed indefinitely.
Funding raised in last 12 months: $0. This is our first external funding request. Six grant applications pending (Solana, Filecoin, Sentient, Coefficient Giving, Alpha-Omega, BlueDot).