The gap
Claiming vs. proving.
What everyone else does
"I know Python" — LinkedIn endorsement from someone who hasn't seen your code
"Built AI systems" — resume bullet, no artifact, no reviewer
"Familiar with LLMs" — nothing shipped, no deployment, no demo
50-day GitHub activity graph — proves consistency, not ability
What Powstik proofs look like
RAG system at [URL] serving 1,200 users — demo recording + architecture diagram, peer-verified by 2 ML engineers who read the codebase
AWS ML Specialty result (Credly badge + exam report) — no self-reporting accepted
Fine-tuned model — eval report + before/after benchmark, endorsed by senior engineer who ran the review
System design delivered to a real team — ADR doc + implementation PR, team lead confirmed it shipped
Why people choose Powstik
Three things no CV can do.
Permanent proof links
Every proof gets a permanent permalink. Share it in a job application, DM it to a recruiter, embed it in your portfolio — the artifact and the endorsements travel with it forever.
Engineer-peer endorsements
The engineers you've worked with verify your proofs directly. Not LinkedIn mutual-endorsements — named people who read your architecture, ran your model, or reviewed your PR.
A corpus that compounds
Every system you ship adds to your verifiable track record. Recruiters can see the depth of your work — not a snapshot, but a history of shipping under pressure.
Top outcomes
Six outcomes with full milestone maps.
Each outcome has 3–6 milestones, each milestone has a specific proof standard and verifier checklist. No self-reporting.
SKILLS · 24 weeks
Become an AI Engineer
6 milestones: RAG system → fine-tuning → production deploy → system design review
SKILLS · 8 weeks
Build your first web app
5 milestones: spec → backend → frontend → deploy → code review
SKILLS · 12 weeks
Master Python
5 milestones: fundamentals → data structures → project → peer review → open source PR
SKILLS · 12 weeks
Master SQL
4 milestones: schema design → complex queries → performance tuning → live project
SKILLS · 16 weeks
Master Statistics
5 milestones: probability → inference → regression → experiment design → ML application
SKILLS · 8 weeks
Earn AWS ML Specialty
3 milestones: study plan → practice exams → certification result (Credly badge required)
Verification
How trust tiers work.
Verification is always human. Named people stake their reputation on your proof — there is no automated scoring or AI quality judgment.
T0 — Self-reported
You submitted the proof. No one has verified it yet. Still permanent and shareable — you're accountable for it.
T1 — Peer verified
One person you've worked with has confirmed the proof is genuine. They signed off using their Powstik account — not anonymously.
T2 — 3+ independent peers
Three or more people have independently verified this proof. The strongest signal a peer network can provide.
T3 — Coach / supervisor verified
A verified coach, academic supervisor, or professional lead has signed off. They hold the verification to the same standard they'd use in a code review or placement report.
T4 — Institutional verifier
The verifier linked their institutional profile (company page, HCPC registration, ORCID, or equivalent) as part of the endorsement. The strongest signal on the platform.
Your corpus starts with one proof.
Log what you shipped this week. Tag the engineer who can verify it. That's the whole loop.
Drop your first proof →