
We just wrapped the very first month of the MLH x DEV Writing Challenge, our new way of celebrating...
We just wrapped the very first month of the MLH x DEV Writing Challenge, our new way of celebrating MLH hackers who take the time to document what they built. We loved reading through the submissions, and we're excited to kick things off by shining a spotlight on our favorite one.
Your hackathon may end after the weekend, but as this month's pick proves, your project story is just getting started.
Let's get to it!
{% card %}
@pritam_patra_429a25dedae6 built OmniML, a tool that helps build AI models and explains what's happening under the hood. Instead of a black box that spits out a finished model, OmniML checks in with you along the way, letting you weigh in on how the model is built, what data it's learning from, and how it's trained. It even explains its own decisions in plain terms and flags whether the result is fair and compliant, so you're never left guessing why it made the choices it did.
{% endcard %}
This month's winner will receive:
Don't let your project story end there. Publish your own submission using the official submission template for a chance to be featured next month.
Hacktoberfest is in full-swing and we're running a new challenge every week, each with a new theme. There's still time to enter this week's challenge:
Register for the upcoming challenges to be notified as details drop:
{% embed https://dev.to/events/challenges/hacktoberfest-week2-2026-10-12 %}
{% embed https://dev.to/events/challenges/hacktoberfest-week3-2026-10-19 %}
{% embed https://dev.to/events/challenges/hacktoberfest-week4-2026-10-26 %}
See you next time!
devchallengeWe are so excited to announce the winners of the Hacktoberfest Weekend Challenge: Build for a Friend,...
discussTL;DR Welcome back to Dev Opportunity Radar. This is a weekly series where I share opportunities,...
gemmaEvery Gemma 4 size from E2B to 31B served on one AMD Instinct MI300X in bf16, fp8, int8 W8A8 and int4 W4A16, on the same vLLM image. fp8 goes from 0.75x bf16 for one request at E2B to 1.23x at 31B and is faster than bf16 in every cell at 12B and 31B. The mixture-of-experts 26B sits at bf16's pace, and the 4-bit builds run at 0.14x to 0.69x at every size.
aiHow replacing fragile linear prompt checklists with a 50-year-old declarative build graph eliminates agent amnesia, enables seamless mid-flight session resumption, and mechanizes software engineering governance.
aiJev is a brilliant decision model. Gemini Flash-Lite is the model nobody talks about, and it’s fast,...
gemmaTen builds of Gemma 4 E2B (bf16, fp8 in both E4M3 flavours, int8 W8A8 and int4 W4A16, each with and without int4 embedding tables) served one after another on one AMD Instinct MI300X with the same vLLM image. fp8 is the only format that keeps pace with bf16, the 4-bit builds run at 0.14x to 0.63x, and int4 embedding tables cost almost nothing.
Workflows from the Neura Market marketplace related to this CoPilot resource