Technical Co-Founder for AI Automation Ventures

Looking for a technical co-founder (equity) — 4 AI / automation projects

I’m building a small portfolio of high-upside products and looking for a hands-on technical co-founder / founding engineer — not a short freelance gig.

What you get
• Equity / profit share on projects that succeed
• Real ownership of architecture, ML, and product decisions
• If none of the projects reach commercial success, a small expense reimbursement (receipts-based stipend; capped — we’ll agree the amount before starting)

What I’m looking for
• Strong Python (ML + production services)
• Comfortable with APIs, bots, and shipping fast
• Bias to ownership over “ticket closing”
• Part-time OK at the start; intensity ramps when something hits traction

THE 4 PROJECTS

────────────────────────────────
1) Polymarket YouTube-views trading bot (live / near-live)
────────────────────────────────
Prediction-market bot that forecasts view-count trajectories on YouTube events listed on Polymarket, then trades bucket outcomes vs the order book.

Timeline: target meaningful live performance (~“15 points” milestone) around Christmas / January.

Technical stack (current):
• Python pipeline: crawl YouTube view series → feature engineering (velocity, acceleration, log views, rolling stats, time-since-start) → ensemble forecast (LightGBM + XGBoost + CatBoost)
• LightGBM tuning focus: grid-search optimal resampling Δt (T_obs / optimal_dt), leaf/learning-rate/regularization, early stopping; compare σ calibration vs XGB/CatBoost; ensemble weighting
• Outputs: μ, σ, P(bucket) via Normal CDF on Polymarket bucket edges; sigma correction + certainty calibration
• Trading layer: EV vs order book, timing (ENTER/HOLD/WAIT/EXIT), Kelly-style bankroll by certainty band; paper + live executor paths
• Ops: Docker, AWS; backtests on paired forecast/poll/trade logs

Ideal co-founder work: LightGBM/ensemble tuning, calibration, risk controls, reliability under live market conditions.

────────────────────────────────
2) Football (soccer) prediction oracle
────────────────────────────────
An “oracle” that produces sealed, timestamped match probabilities before kickoff — usable for tipsters, B2B feeds, or on-chain/off-chain settlement-style trust (integrity > marketing).

Technical direction:
• Match outcome / score models (e.g. Poisson / Dixon–Coles or modern GBMs + Elo/strength ratings)
• Features: team form, home/away, injuries, rest days, league strength, market odds as prior
• Calibration (Brier, log-loss), probability → fair odds; optional Kelly / staking layer
• Immutable “sealed before kickoff” ledger (hash + timestamp) so tips can be audited
• APIs + simple dashboard for tipsters / partners; clear “out of sample / insufficient data” flags

Ideal co-founder work: model rigor, calibration, data pipelines, and a product that tipsters/partners actually trust.

────────────────────────────────
3) Restaurant / hospitality booking bot (highest commercial upside)
────────────────────────────────
Automation product for restaurants (and similar venues): reservations, waitlists, confirmations, reminders — WhatsApp/Telegram/voice — with a clear SaaS / per-booking or subscription business model.

Why this matters: solid unit economics; with early traction this is the most VC-backable of the four.

Technical direction:
• Conversational booking flows (chat + optional voice via Telnyx/Twilio)
• Calendar / table inventory, no-show reduction, reminders, multi-language
• Integrations: POS / Google Calendar / venue CRM; onboarding + admin panel
• Analytics: fill rate, no-shows, revenue recovered

Ideal co-founder work: product + integrations + reliability; help turn pilots into paid traction.

────────────────────────────────
4) US wholesale real estate outbound phone bot
────────────────────────────────
Voice AI / dialer for US wholesale RE: qualify sellers/buyers, leave compliant messages, book appointments, push leads into CRM.

Technical direction:
• Voice stack (Twilio / Telnyx + STT/TTS + LLM dialogue)
• Parallel dialing, disposition codes, DNC / TCPA-aware calling windows
• CRM sync (GoHighLevel, HubSpot, or custom), SMS follow-up
• Scripts for motivated sellers, absentee owners, buyers; appointment booking

Ideal co-founder work: voice pipeline reliability, compliance-aware flows, and conversion metrics.

────────────────────────────────

How we’ll work
• Async-first; English or Italian OK
• Start by picking 1–2 projects to own; expand if traction appears
• IP and equity terms documented before deep work

To apply, please answer:
1) Which project would you take first, and why?
2) Link to relevant work (GitHub, product, or short write-up — LightGBM / bots / voice preferred)
3) Hours/week available for the next 90 days
4) Confirm you’re OK with equity + expense reimbursement only if nothing commercializes

Serious builders only. If you want a fixed-price “build my app” job, this isn’t it.