Ash Bhatta

It s easy to build in a bubble. It s safe. You and your team can convince each other that everything is working exactly as planned.
That s exactly where we were before we launched CareerButler in April.
But once we opened the doors to 400+ active users, the bubble popped. We had to separate the noise from the signal.
We learned that validation isn't a checkbox you tick once. It s a constant loop.
Build what you think works.
Get hit with reality (feedback).
Evolve.
Repeat.
We realized that a tool isn't useful because we say it is. It's useful when a frustrated job seeker lands an interview, or a recruiter finally sees a readable resume.
We moved from Assumption-Based Development to Evidence-Based Development.
That evidence is the core of CareerButler V2.
We are launching this Sunday on Product Hunt. Come see what happens when you actually listen to the market.
Check out the teaser: https://lnkd.in/eJgq7Bqi

Here is the issue we ran into while building CareerButler V2:
We built this powerful AI to tailor resumes to job descriptions. But we noticed that if the user uploaded a weak resume to start with, the AI couldn't save it.
If your resume says "I led a team to success" without any numbers, the AI doesn't know what 'success' means. It can't optimize what isn't there.
So it either:
Polishes the fluff (still useless).
Starts hallucinating and making up numbers (dangerous).
We realized we couldn't just build a 'tailor'. We had to build a 'fixer' first.
It's called Resume Critique.
Before we begin optimizing your resume for a specific job, we force a deep scan. It parses your resume section-by-section to fix the foundation.
We actually highlight the specific bullet points that are weak, right on the screen, and give you tips on exactly how to improve them.
And I don't mean generic advice like 'Add more metrics.' That is lazy.
I mean specific, actionable pushes.
Here is the difference between Generic Advice vs. CareerButler V2:
Generic Tool: "You should add more metrics."
CareerButler: "In your role as Project Manager, you listed 'Led a team to success.' This is vague. Consider adding the team size and a specific outcome. For example: 'Led a team of 10 engineers to deliver the product 2 weeks ahead of schedule.'"
Generic Tool: "Your summary is too long."
CareerButler: "Your professional summary is currently 6 sentences long. Recruiter heatmap research shows they stop reading after sentence 2. We recommend cutting this down to a highlight reel of your top 3 hard skills."
It s like having a career coach proofread your resume before you start applying. It stops you from optimizing a bad resume.
Fix the foundation first. Then tailor it.
We are launching V2 on Product Hunt this Sunday. Come see if your resume passes the check.

We get asked this constantly. It s a fair question.
If you want to write a poem or a recipe, use ChatGPT. It is an incredible Generalist. But if you want to navigate a complex, rigid system like hiring, you don't need a generalist. You need a Specialist.
Here is the deep-dive on why we built CareerButler despite the noise:
1. The Prompt Engineering Gap
Yes, ChatGPT can write resumes. But the output is only as good as your prompt. Most job seekers aren't prompt engineers. They ask for a resume, and the LLM gives them creative writing, fancy adjectives and hallucinations.
Granted ChatGPT is getting smarter by the day. But raw intelligence isn't the issue, context is.We didn't build CareerButler to compete with AI; we built it to make AI work specifically for resumes.
2. The Competitor Flaw. They're Literally Just Paying ChatGPT
Honestly, this is the part that drives me nuts. Most of the paid tools you see are just copy-and-paste apps. They have a nice website, but all they do is take your resume, send it to the same AI everyone else uses, and then slap a price tag on the result. They're not adding any real smarts or unique engineering.
Because they aren't actually solving the problem, they resort to gimmicks like:
- The False ATS Score: This is the biggest scam in the industry. We found tools using what seemed like random number generators to give you a compatibility score. It s designed to make you anxious so you pay them, not to help you get hired.
We don't give you a black-box number. We visualize the why. If your score goes up, we show you exactly which keyword or formatting change caused it based on reverse-engineered logic from systems like Greenhouse and Workday.
3. Structure vs. Content
A free text generator can t fix your document structure. It can write a bullet point, but it can t ensure your margins, date formats, and hierarchy are readable by an older parsing algorithm.
We built a hybrid engine. We use LLMs for the content, and we intelligently fit the new optimized resume into a downloadable file, ensuring proper formatting, layout, and number of pages depending on your experience. This ensures the file structure is perfectly machine readable and the text is optimized for recruiters
See the difference between a Generalist and a Specialist.
We are launching V2 on Product Hunt this Sunday: CareerButler