northbound-capital · posted 2h ago
Applied AI Engineer · Finance
$50–70k · remote-friendly · IC3
→ skills extracted
[ the new resume for the self-taught AI generation ]
Finish a challenge.
Unlock 100 jobs.
Every Proof challenge is reverse-engineered from a real job description. Ship it, your AIQ + Proof score moves, and the open roles you just qualified for appear. No CV. No applying.
[ The loop — this is the whole product ]
Every challenge is a job. Every job is a challenge.
Job description
Companies post real open roles. We tag the skills, the stack, the seniority.
Challenge
Each JD becomes a timed challenge. Build the thing the job actually requires.
Score moves
AIQ + Proof recompute the moment you ship. Cumulative. Public. Portable.
100 jobs unlocked
Every JD that maps to your new score appears, ranked by match. One-click intro.
→ Companies don't post jobs to a board. They post challenges. The winners are the hires.
[ Live challenges ]
Every challenge is wired to real jobs, real money, real seats.
New drops every weekend, sourced from JDs companies posted that week. Sprint in 48h or go deep in two weeks. Top 5% gets a verification call.
See all live challenges →[ The moment after submit ]
Prooft the leaderboard. Show me the jobs.
You just unlocked
137 jobs
AIQ 82 → 90 · Proof 79 → 91 · +20 in one weekend
Northbound Capital
Applied AI Eng · Finance
Lex.ai
Founding AI Engineer
Vega Health
AI Eng · Clinical RAG
Shibuya Labs
Agent Eng · Tools
Mint & Co
AI Eng · Tax Automation
This is the screenshot people send their parents. The moment a self-taught builder stops being invisible and becomes pre-qualified for 137 open seats — without ever opening Naukri.
[ The two scores · cumulative, not snapshot ]
Two numbers replace a hundred bullet points. Both move every weekend.
↗ +18 in last 90 days
AI fluency. How well they build with models, ship agents, design prompts and evals, and deploy real systems. Weighted heavily by live challenge performance — hard to fake.
- +Live challenge results (heaviest)
- +Working with LLMs & agents
- +Evals, fine-tuning, deployment
- +Peer + employer endorsements
↗ +12 in last 90 days
The composite. Everything they've shipped, across every domain — law, finance, design, engineering — judged on outcome, recency, and range. Decays slowly so you can't coast.
- +Shipped projects & GitHub depth
- +Challenge submissions (cumulative)
- +Domain depth × cross-domain range
- +Outcomes, not credentials
→ Score = Σ (challenge_score × difficulty_weight × recency_decay). Builders are training, not applying.
[ The artifact — one profile, one summary ]
The new resume is a score, a body of work, and a summary written for the hiring partner.
VERIFIEDAlex Kim
AI Engineer & Builder
- San Francisco, CA
- 5+ years experience
- M.S. Computer Science, Stanford
- alexkim.com @alexkim
AIQ
94
Measures AI problem solving, system design & intelligence
Proof
87
Measures real-world impact, execution & outcomes
SELECTED PROJECTS
TERRAIN
3D generative AI
AGENT OS
multi-agent framework
ECHO
real-time intelligence
DATAWEAVE
data pipeline optimizer
LUMEN
text-to-3D engine
NOMAD
AI travel copilot
SUMMARY FOR HIRING PARTNER
Alex is a builder who turns complex AI concepts into scalable, production-ready systems.
Combines strong technical depth with a design sensibility and a clear bias for impact.
Proven track record of shipping products that drive measurable outcomes.
Highly autonomous, collaborative, and energized by hard problems.
One link. Two scores. Six things they shipped. The summary is auto-written for the role you're hiring for.
[ For employers ]
Post a JD. We turn it into a challenge. Hire the people who actually did the job.
Stop reading 5,000 PDFs. You give us the role; we ship a challenge in a week; builders compete; you receive a ranked shortlist whose submission is their application.
Top submission · auto-summary
Alex ranked #3 of 412 on your "contract review agent" challenge. AIQ 94 · Proof 87, +18 in last 90 days. Shipped 14 production agents, won 3 challenges in legal & finance.
jd.txt — pasted just now
[ For universities ]
We run your placements. All of them.
Every student gets a Proof profile from day one. By final year they've shipped 200+ challenges and graduated pre-qualified for hundreds of roles. Companies hire straight from your campus pool. We charge a retainer.
One profile per student
AIQ + Proof, built across 4 years from coursework, projects, internships and challenge submissions.
One pool per campus
Companies search your pool by score, by domain, by challenge they ran — and request intros directly.
One retainer per university
Proof runs as your placement infrastructure. Annual contract, white-label optional.
0
students onboarded
[ For freshers ]
Start in your 1st year.
Graduate hireable.
A score that grows with you. Do challenges every weekend through college. By the time placement season hits, you have four years of shipped work, not one frantic semester of LeetCode.
Year 1
Proof 18
First weekend challenge. Score on the board.
Year 2
Proof 41
30 challenges in. First internship intros land.
Year 3
Proof 68
Domain depth shows. Companies start reaching out.
Year 4
Proof 84
Graduates pre-qualified for 400+ roles.
→ Internships go to whoever cleared the internship-tier challenge. No interview required.
How it works
Pick a challenge
Reverse-engineered from a real JD. Pick by domain, by stack, by salary band.
Ship + get scored
Submit repo + deploy URL + 60s loom. AIQ + Proof recompute live.
Unlock the jobs
Every JD that maps to your new score appears, ranked. One click sends your submission.
[ Built for both sides of the new hire ]
A vibe coder belongs in your CA firm. An AI builder belongs in your law firm.
[ Manifesto ]
Job boards collect resumes. We collect proof.
Every other hiring platform asks you to apply. We ask you to build. Every challenge you finish moves your score and unlocks the next 100 jobs that needed exactly that skill. The self-taught generation doesn't need a CV — they need a place where the work itself does the talking, and the moment they ship, the door opens.