Steve Jobs : how to design perfect products.

How to make Cursor love your monorepo.

Day 24 of building in public Not all progress is visible. Some days it’s not a launch, a feature, or a shiny update. It’s: • Refactoring messy code • Rethinking user flow • Rewriting a line 10 times • Talking to a confused user Quiet progress still counts. Keep building.

Day 23 of building in public It’s easy to chase the “big” moves a launch, a redesign, a new feature. But in reality? Most progress happens in the small stuff: • Better onboarding • Faster loading • Clearer copy • Smoother UI Tiny changes. Big impact. At Masar, that’s

Day 22 of building in public Some days, I look around and feel late. Late to AI. Late to startup life. Late to success. But then I remind myself: There are no late bloomers just people who kept going. I’m building Masar on my own timeline. And that’s enough.

Day 21 of building in public Not every piece of feedback deserves action. That was a hard lesson. At first, I tried to fix everything users mentioned until the product lost its direction. Now? I listen to everyone… But only act on what aligns with our vision. Feedback ≠

Day 20 of building in public There’s no fancy team. No office. No funding. Just two founders building day and night, writing code, fixing bugs, talking to users, rewriting copy, designing pages... It’s messy. It’s hard. It’s real. And we’re showing up every day until it

Day 19 of building in public Some days, you’re on fire. Other days… it’s just you, your screen, and a quiet “why am I even doing this?” That’s when the real work begins. Showing up when it’s not exciting, that’s what separates hobby from mission. And I’m choosing to show

Day 18 of building in public Move fast, but not so fast you break trust. At Masar, we’ve had to make painful tradeoffs: Shipping fast vs building stable, polished experiences. We chose this principle: “Ship lean, polish in public.” It’s not perfect, but it gets us in motion

Day 17 of building in public Nothing is harder than rewriting something you already built. But nothing is more necessary when you know it won't scale. We recently refactored a big part of Masar, not because it was broken, but because we’re building for what’s coming, not what

Day 16 of building in public One of the hardest early decisions we made: We scrapped a “cool” feature that students loved in testing… Why? It didn’t align with our core promise: helping them choose the right major , fast, focused, and free of distractions. It hurt, but

Day 15 of building in public One lesson I keep relearning: The longer you hesitate, the heavier the decision feels. In startups, speed isn’t just about moving fast . it’s about reducing the weight of “what ifs” before they pile up. Decide, test, adjust. Repeat.

Day 14 of building in public One of the first tough calls we made: Not to use generic chatbots. We wanted every student to feel like they’re talking to someone who gets them, not just a form with AI sprinkles. So we built a fully custom AI experience tailored for students.

Day 13 of building in public One thing nobody tells you: Most of the hardest decisions happen before anyone sees the product. Like: Should we build our own AI experience, or plug into existing tools? How much friction is “too much” when guiding students? Should we launch

Day 12 of building in public Watching students use Masar for the first time is… something else. Some get excited. Some feel seen for the first time. Some even say: “This feels like it was written for me” That’s when you know, it’s not just code anymore. It's real impact

Day 11 of building in public Some of the hardest decisions in building a startup… never make it to the roadmap. Things like: – What features to delay, even if they’re exciting ? – How much magic is too much before it feels fake ? – When to simplify, and when to double down ?

Day 10 of building in public Masar isn’t just a quiz. It’s an AI model built to understand students, their strengths, tone, and hidden potential. We trained it to act like a mentor, not a machine. The dream? That every student finds clarity, confidence, and direction — in less

Day 9 of building in public AI doesn’t have to feel cold or robotic. When building Masar, we focused on making the experience feel like a friendly conversation, not a test. Students should feel seen, not scanned. It’s one of the hardest UX challenges we’ve faced. But the

Fine-tune 100+ LLMs directly from a UI! LLaMA-Factory lets you train and fine-tune open-source LLMs and VLMs without writing any code. Supports 100+ models, multimodal fine-tuning, PPO, DPO, experiment tracking, and much more! 100% open-source with 50k stars!

Day 8 of building in public We quietly passed 1,000+ students trying Masar That’s 1,000+ unique journeys. 1,000+ decisions that now feel a little less confusing. The most common feedback? “It felt like someone finally understood me” That’s why we’re building. You can try it