mujeeb

Saw the Claude Certified Architect exam and decided to build a curriculum around it. The idea: design a production-grade project that forces me to learn every concept deeply, then build it. Took about 3 weeks alongside my day job.

@_chenglou 😂

Years of what now

@Tancrededib Kinda know the right person for this, insanely resourceful and curious, I’ll send him this tweet

Looking for someone early in their career with strong taste in people, high curiosity, and real excitement for company building. A lot of the role is: find exceptional people and help the best go build in San Francisco.

I’m hiring someone who: - has taste - loves people - enjoys LDN & SF - wants to shape the future DMs are open

Unlocked the ultimate teacher

😂

@levilian1 AI-generated summaries from PubMed abstracts, seeded with figures from SUSTAIN-6, SELECT, LEADER etc. The agent reasoned over structured synthetic data, not raw PDFs. The limitation is provenance: I was the curation step between the real source and the eval cases.

@__mujeeb__ Exactly! Not to mention labs buy data from bench providers and post train… What datasets did you use?

(1/7) Up until recently, we’ve treated LLMs like a magic 8-ball. You type a prompt, shake it, and hope a correct answer floats to the surface. If you draft emails, that works. But in law, consulting, or biopharma, a plausible text string isn't enough. 🧵👇

Trained in a Nigerian tertiary hospital with no EHR. The admin burden was still massive, handwritten case notes, paper investigation requests, manual discharge summaries. The technology wasn't the cause. The documentation requirement itself was. Burnout without a screen to blame it on.

Real burnout numbers: • 63% of US physicians report burnout (AMA 2025) • #1 cause: Administrative work (not clinical complexity) • Average doc spends 2 hours on EHR for every 1 hour with patients Fix *that* and you print money. (3/7)

Healthcare AI cos are solving the wrong problem. Clinicians don't need better diagnoses. They need time back. Here's the $100B insight: (1/7)

This is the exact problem with LLM-as-judge evals. The judge scores the output, not the consequence. I built an adversarial eval tier for a clinical trial agent specifically because passing 28/30 cases meant nothing if the 2 failures were confident wrong citations. The benchmark hides what matters.

(2/7) I've spent the last year looking at why frontier models still struggle in enterprise deployment. Top agents pass rigorous benchmarks like APEX 30% of the time. But that score hides a bigger problem: an agent can pass while generating a document that gets an associate fired.

@dereckwpaul The citation layer is what makes or breaks clinical AI in practice. Confident answers with wrong sources are worse than no answer. Curious how you handle retrieval failures when the query doesn't map cleanly to any guideline.

Evidence-based clinical intelligence is becoming a critical infrastructure for many healthcare technology companies. We've made building with our clinical AI agent exceptionally easy with in-app self-serve access to the Glass Developer API. Developers can get started building with our AI today and bring new evidence-based diagnostic, treatment planning, and documentation capabilities to their products and platforms.

@_ayoobami so strict yo, I forgot I set it to be straight like that with me

@__mujeeb__ Bro said enough bro ,go and read

chill bro 😂😂

I want to found an AI eval framework for healthcare. B2B. Eval tooling today is generic. Healthcare needs domain-specific evaluation where clinical accuracy matters. I'm an MD who coded through 6 years of med school. I understand both sides. 2 production agent systems with eval pipelines built. Currently at Mono (YC W21). Lots of energy and agency to burn.

@__mujeeb__ And you want to found ?

Investing $250k in non-US based builders. 40 spots, 8 weeks, San Francisco. Flights/housing/food covered + visa support. This Summer Looking for: technical/product builders pre/early idea cofounder-seeking Comment your edge, I’ll get back to you!

@tokifyi Broke the matrix 😂

you can just do things 😂

@PavanKumarNY Heyo

If you are building in the AI agent space, I want to talk to you. Let’s connect 🤝

@gao_zibo You don’t even gotta get told this, you’ll find out

Stop building disposable apps Build for permanence

@shauseth actually, I have a feeling thing will be somewhat different if he was around

god i miss steve jobs. we are truly lost

Yup, I was thinking along these lines today. I’ve been learning a lot with AI over the past few months, and I genuinely think it’s the best way. A personalized curriculum: after every module, I’m asked detailed, first-principles-esque questions about what I’ve learned, and I’m corrected on my shortcomings. Detailed notes are created after major phases All of this happens while I’m building an actual project. It’s so refreshing.

Been thinking about this: A SaaS that auto-generates learning roadmaps. Not generic ones. Based on what's actually working right now. You want to learn AI engineering? It shows you what people are shipping today. Not what was hot 6 months ago. The roadmap updates itself. Because trends move faster than any course can keep up.

The biggest change AI has done for me is how I learn new things, its incredible how much you can learn if you do it right

@krishdotdev +1

Twitter is cool but it’s 10x better when you #connect with people in your niche. If you’re into: > LLMs + RAG > AI Agents > AI Automation > System Design > Product Thinking > AI Driven Development > OpenClaw automations > SaaS & Product distribution Drop a +1 & let’s connect.

@tokifyi motion

update: scheduled 3 vc calls for tomorrow