frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: LiminalML – Study ML or SWE at interview depth, grounded in your resume

https://liminalml.com
3•OliverPerrin•1h ago

Comments

OliverPerrin•1h ago
I built this because when I was prepping for interviews I had lot's of separate material in different places I was referencing to study. I was bouncing between Claude/GPT sessions that would lose some context, Stanford lecture notes, exam questions I'd collected, and my own notes. None of the material really connected together much and it made it hard for some things to stick.

LiminalML goes through 157+ ML (or SWE) topics in six stages that build on each other, ending with a retrieval check grounded in your resume if you choose to upload it, allowing the practice questions to reference your projects or work experience. It renders LaTeX nicely and code inline without prompting and keeps context across your whole prep cycle.

It isn't strictly a coding practice either, tools like Deep-ML do that well. This is for the conceptual and theory side, which I find often gets a bit ignored by Leetcode style websites, so I wanted to avoid that style. You get walked through an overview, intuition (with visuals), math, code, interview questions, and a retrieval check at the end as mentioned above.

Free tier has 8 sessions/month. Pro ($9/month) adds a practice lab where you can code while referencing your sessions as well as unlimited sessions. I'm still in the early stages of building so feedback on what's missing or broken is greatly appreciated and will go a long way.

SF enters its building era, kicking and screaming: SF Standard Podcast

https://www.youtube.com/watch?v=k4KOd6e4_-I
1•scoofy•2m ago•0 comments

Ask HN: Who Has This Pain?

1•swk-phil•3m ago•0 comments

iMessage is now possible on Android (again)

https://www.neowin.net/news/this-app-promises-to-bring-imessage-to-androidagain/
1•bundie•5m ago•0 comments

Show HN: Learning Nix with LabCraft

https://www.labcraft.dev/
1•anandsuresh•5m ago•0 comments

London cops handed victim's new address and number to her stalker, watchdog says

https://www.theregister.com/security/2026/08/05/london-cops-handed-victims-new-address-and-number...
4•Bender•7m ago•0 comments

Weeks into explosive diarrhea outbreak, sluggish CDC plans response team

https://arstechnica.com/health/2026/08/report-cdc-finally-plans-formal-response-team-for-explosiv...
5•Bender•8m ago•0 comments

Apps for Kobo

https://github.com/BandarLabs/cobalt
1•mkagenius•8m ago•0 comments

What happens if you put work into the second dimension?

https://norbertkozsir.com/posts/work-in-the-second-dimension/
1•abelsm•8m ago•0 comments

A Harness for AGI

https://viktor.com/research/a-harness-for-agi
1•everlier•9m ago•0 comments

Hey AI come train on this song [video]

https://www.youtube.com/watch?v=ZBvLadXo7XA
1•haolez•9m ago•0 comments

Born Against, or why hobby programming communities are against LLM usage

https://blog.fogus.me/llm/born-against.html
3•lladnar•10m ago•0 comments

OpenFlexure Microscope – The OpenFlexure Project

https://openflexure.org/
1•latchkey•10m ago•0 comments

Show HN: DeepSeek V4 Flash 0731 with MoonViT Vision (NVFP4)

https://huggingface.co/webbrain-one/DeepSeek-V4-Flash-0731-Vision-NVFP4
1•webbrain•12m ago•0 comments

FlockHopper

https://dontgetflocked.com/
1•cdrnsf•12m ago•1 comments

Why low-latency trading infrastructure is harder than it looks

https://medium.com/@yaroslavaristov/why-low-latency-trading-infrastructure-is-harder-than-it-look...
1•yaroslavaristov•14m ago•0 comments

Can Agents Deceive? Evaluating Reasoning+Deception Using a Social Deduction Game

https://arxiv.org/abs/2607.28146
1•theanonymousone•14m ago•0 comments

The Lua community needs to learn to move on

https://hisham.hm/2026/08/04/the-lua-community-needs-to-learn-to-move-on/
3•thunderbong•15m ago•0 comments

Craft a detailed prompt to support test activities

https://club.ministryoftesting.com/t/day-8-craft-a-detailed-prompt-to-support-test-activities/74863
1•mahirsaid•17m ago•0 comments

hbkit: Recover Synology Hyper Backup (.hbk) archive files

https://github.com/YordiLorenzo/hbkit
1•toomuchtodo•17m ago•1 comments

New Drug: AI

https://blabbingabout.substack.com/p/new-drug-ai
1•atkntepe•18m ago•0 comments

Deployah – deploy to Kubernetes from a short spec, no Helm, nothing in-cluster

https://github.com/deployah-dev/deployah
2•atkrad•21m ago•2 comments

A fake Galaxy S22 Ultra, taken apart layer by layer

https://chardonnet.github.io/fake-s22-ultra-teardown/
2•techburger•22m ago•0 comments

DeepSight – give text-only LLMs eyes and hands (zero tokens, on-device)

https://github.com/Reality-Shifting-Tech/deepsight
1•eliaseffects•22m ago•0 comments

Show HN: GetPageSpeed Amplify, a compatible replacement for Nginx Amplify

https://amplify.getpagespeed.com/
1•dvershinin•23m ago•0 comments

Tracking Apple's Environmental Claims Across Product Generations

https://tostracker.app/entity/apple-inc/environment
1•tldrthelaw•27m ago•0 comments

Album Whale

https://albumwhale.com/
2•pentagrama•28m ago•0 comments

(Updated) Project Phoenix- Autonomous AI Agents

https://phoenix-production-96cd.up.railway.app/
2•amonte•28m ago•0 comments

Show HN: Hacker News with reduced priority for AI driven content

https://sprinklz.io/public/nncrnx74ykir
2•sammy0910•29m ago•0 comments

Microsoft's AI Sales Mostly Come from OpenAI, Disclosures Show

https://www.bloomberg.com/news/articles/2026-08-05/microsoft-s-ai-sales-mostly-come-from-openai-d...
8•mapping365•29m ago•2 comments

Beating GPT-5.6 Sol on retrieval with 100x cheaper open models

https://neon.com/blog/how-castform-neon-beats-frontier-models-on-price-and-efficiency
4•moonikakiss•29m ago•0 comments