Devpost
Participate in our public hackathons
Devpost for Teams
Access your company's private hackathons
Grow your developer ecosystem and promote your platform
Drive innovation, collaboration, and retention within your organization
By use case
Blog
Insights into hackathon planning and participation
Customer stories
Inspiration from peers and other industry leaders
Planning guides
Best practices for planning online and in-person hackathons
Webinars & events
Upcoming events and on-demand recordings
Help desk
Common questions and support documentation
Sort by:
When hospital rooms go dark, DualPulse AI watches over beds. Powered by OpenCV 5 and AWS Graviton4, it catches real patient falls in real time with alarms.
AutoScope turns a smartphone and a standard microscope into an adaptive malaria screening tool. It counts parasites automatically and adjusts how much of a slide to read based on the evidence.
An intelligent study workspace leveraging OpenCV and AWS to analyze focus, enhance student engagement, and provide interactive real-time visual learning assistance.
What's the next move? How do we solve it? A local friendly tutor so you can learn to solve Rubik's cube.
Offline data-center parts identifier on the Alif E8: U85 NPU names the parts in 7.9 ms, voice opens its checklist, specs & troubleshooting on device. Ext, Int service at DC simplified for tech.
DBX is a high-performance, multi-tenant vector database built in Go. Engineered with SIMD-accelerated search and memory-mapped persistence for ultra-low latency AI embeddings retrieval at scale.
Machine vibration diagnosis from ordinary handheld video. A real ceiling fan read at 144 RPM, no contact sensor. OpenCV 5 plus an agent that decides what to film next.
An autonomous robotic arm that can interpret, discuss, and execute physical tasks at your command.
Agri-Bot is an agricultural robot that helps in several ways in agriculture: it detects pests and gets rid of them, checks the temperature and water levels of the crops, and many other features.
SmartSort AI is a Gemini-powered intelligent waste-sorting robot that combines multimodal AI perception, reasoning, and robotic actuation to automatically identify and sort recyclable materials
QuantForge runs a real optimization experiment directly on an Arm-powered Android phone.
Hand-written SDOT/SMMLA kernels make quantized LLM inference up to 3.2x faster on Arm — bit-exact with the baseline, and every number reproducible on free Arm64 CI without owning Arm hardware.
Energy optimization for ML/AI workloads for ARM cloud servers
On-device agent harness optimized for Arm: KleidiAI-tuned 1.2B on a $150 Galaxy M17, 34x faster first token, plus close-to-hardware shell access root-capable, sandbox-aware device control. Zero cloud.
The attribution ladder for Arm inference builds the real unaccelerated floor, splits the speedup by mechanism, and proves which kernels actually ran.
Measure the Arm machine you're actually on. GraviTune found that llama.cpp's default setting costs 2x prefill throughput on AWS Graviton4 proven with hardware counters and filed upstream.
Phase-Aware CPU thread autotuning for llama.cpp on Arm Android. Solves big.LITTLE barrier lock contention to deliver a verified 2.07x sustained decode speedup and 9.4x latency stability.
On-device gesture control system, optimized end-to-end for Arm architecture.
LLM inference for the other 6 billion. No GPU. No cloud. Just ARM.
Find out which Arm CPU path your AI inference actually took — and what it cost in joules per token.
This project involved developing and benchmarking 3 improved ARM kernels for Gated Delta Net, a linear-complexity replacement for transformers used in the latest Qwen, Nemotron and Kimi models.
Ask your lecture slides a question and get an answer with the slide number — offline, on a 2020 budget phone. 3x faster to first word, and none of it came from Arm's KleidiAI. We measured.
On-device PII redaction for Arm: a from-scratch int8 transformer in C with five hand-written AArch64 kernel tiers and runtime dispatch that never picks an instruction your silicon can't execute.
Cost/SLO-aware LLM serving optimizer for AWS Graviton4 (Arm Neoverse V2) with continuous batching, KleidiAI INT4, Arm Performix PMU hardware proof, SHA-256 ledger, and MCP tool interface.
