Frank @ Four-Leaf

Founder of job search assistant built around you. Voice interviews, resume tailoring, comp coaching, smarter search.

[Frank @ Four-Leaf @FourLeaf_AI]

We're excited to share that we shipped a Four-Leaf MCP today. 11 tools for AI job search and interview prep, accessible inside Claude, Cursor, ChatGPT, Perplexity. All of the Four-Leaf tools now available in your AI assistant

You can find more details at: https://t.co/mAhInFNEIz

[Frank @ Four-Leaf @FourLeaf_AI]

76.4% of recruiters rank candidates by the exact skills in the job description.

Most systems match your words, not your intent. "ML engineer" and "machine learning engineer" read as two different tokens. Mirror the posting's wording, don't paraphrase it.

https://t.co/EGcj9PLHaw

[Frank @ Four-Leaf @FourLeaf_AI]

The labs everyone wants to work for tell you what belongs at the top of your resume, and it isn't your degree.

Anthropic's careers page says to lead with independent research, a blog post, or open-source work. Proof of work beats a credential list.

Put the proof at the top.

[Frank @ Four-Leaf @FourLeaf_AI]

@e_opore We're running #2 in production at https://t.co/fhFQYkrR4e (not Java, but same shape). The hardest part wasn't the matching model, it was building a JD canonicalizer that handled the 30 different ways companies write the same requirement.

[Frank @ Four-Leaf @FourLeaf_AI]

@heyblake Distribution

[Frank @ Four-Leaf @FourLeaf_AI]

@Shay_Slay_ We built #1 (AI interview simulator with voice and adaptive follow-ups) and #2 (resume analyzer scored against the JD) at https://t.co/fhFQYkrR4e. Happy to compare notes with anyone building either.

[Frank @ Four-Leaf @FourLeaf_AI]

True for any agent, but it's the whole game for a coaching agent. Without the JD, the resume, and the candidate's actual stories loaded in, you get generic STAR advice. With them loaded, you get rehearsal that actually transfers.

[Aaron Levie @levie]

The main variable in getting success with agents is whether you can get the agent the context it needs to do its work; and a major factor in that is if you can create a shared working area for that agent that a human can understand as well.

This is one of the reasons why agents using file systems is such a big deal. It creates a unified system that both the person and the agent can work within to pass around data.

“What they need is a working set: plans, notes, task lists, policies, drafts, summaries, logs, corrections, decisions, etc. For that layer, a filesystem-shaped interface tends to be more legible to both the model and the humans supervising it.”

It turns out giving agents access to the systems we already know how to use, but in a way that is best optimized for them, is the perfect primitive for agents to work.

[Frank @ Four-Leaf @FourLeaf_AI]

Same pattern on the interview coaching side. The candidates who used to book $300 sessions are now using AI coaches. Coaches who survive will need to be doing strategy and emotional work that the models can't do as well

[Lenny Rachitsky @lennysan]

Heard from an executive coach that her business is way down due to AI.

[Frank @ Four-Leaf @FourLeaf_AI]

'Where do you see yourself in five years' lands worse than it sounds, and most answer it like a trap.

The weak version recites a title ladder. The strong one names a skill they want to get great at and ties it to the work this role involves.

Name the skill, not the title.

[Frank @ Four-Leaf @FourLeaf_AI]

Ten minutes before an interview, skip re-reading your resume.

Pull up the company's latest news so you can name one specific thing. Pick the two job requirements your stories will hit. Then say your opening line out loud.

A specific reference beats a silent read-through.

[Frank @ Four-Leaf @FourLeaf_AI]

@samfbauer Tour guide

[Frank @ Four-Leaf @FourLeaf_AI]

Only about 1 in 50 job postings names a specific AI tool like Copilot or ChatGPT, even though nearly 1 in 4 mention AI (Four-Leaf's analysis of 210,000 open postings).

Most "AI experience" lines are vibes, not a named skill. Prep a story for them like any behavioral signal.

[Frank @ Four-Leaf @FourLeaf_AI]

When I ask your salary expectations early, I'm not trying to trap you.

The answer that helps names a researched range tied to market and role. Not "I'm flexible," and not a number so low I wonder what you think the job is.

A range with a reason reads as knowing your worth.

[Frank @ Four-Leaf @FourLeaf_AI]

The job is filled. The posting is still up. I can explain both.

https://t.co/RN5M1IXVUu

[Frank @ Four-Leaf @FourLeaf_AI]

Coinbase cut 14% of staff this month, and CEO Brian Armstrong tied it to AI changing how the company works (per Crunchbase).

These are restructurings, not downturn cuts. Roles lost this way rarely come back. Weight the skills these teams keep.

https://t.co/IpSTCMJQhN

[Frank @ Four-Leaf @FourLeaf_AI]

@Joi2James Four-Leaf is the job search assistant that builds real confidence through real prep. Not with shortcuts or AI overlays that get candidates blacklisted. Just simple and holistic prep that helps candidates breakthrough. Let’s connect

https://t.co/nj36wKMeeK

[Frank @ Four-Leaf @FourLeaf_AI]

@TTrimoreau Real Prepforcanadiatesthathelpsthemeverystepoftheway

https://t.co/nj36wKMeeK

[Frank @ Four-Leaf @FourLeaf_AI]

@tinytechfox @saas_core Four-Leaf is the job search assistant that builds real confidence through real prep. Not with shortcuts or AI overlays that get candidates blacklisted. Just simple and holistic prep that helps candidates breakthrough.

https://t.co/nj36wKMeeK

[Frank @ Four-Leaf @FourLeaf_AI]

@JunaidAckroyd Focus on distribution like I should be doing anyway… product is great already, but coding is so much more fun than outreach and content

[Frank @ Four-Leaf @FourLeaf_AI]

@zdogmode https://t.co/F68AtKA5BG

[Frank @ Four-Leaf @FourLeaf_AI]

@boardyai Ice cold take Boardy