Project Description
EdgeStack is a professional trading workstation that combines a proprietary technical signal engine (VZO + EMA60 + ADX, ported verbatim from a battle-tested Python core) with an AI-grounded assistant that reads live cached market data via Claude tool-use. Today's stack: FastAPI + SQLAlchemy async backend, React + TypeScript + Vite frontend, Polygon WebSocket for live market bars, CBOE for ^VIX, CoinGecko for crypto, RSS feeds for news. The system already runs end-to-end with 18 headline ETF cards, a deep-dive scanner with 9 modes, multi-advisor AI consensus, option chain + ThinkorSwim order generation, watchlist + holdings + trade log persisted client-side, and a 6-section Settings panel.
The product is real and demonstrable — what is missing is the path from "working prototype" to "shippable SaaS." That is what this engagement is about.
A senior full-stack engineer who has shipped a real consumer or pro-sumer SaaS before — ideally in fintech, but strong B2B SaaS experience also welcome. You should be comfortable owning architectural decisions, not just executing tickets, and willing to push back when you see something that won't scale.
This is a long-term collaboration. I am the product owner and primary stakeholder; you would be the technical partner. Expected commitment: 20-40 hours per week, 3+ months minimum, with the option to extend.
Please address all three. Generic proposals copy-pasted across listings will be ignored.
1. Persistence & multi-device sync architecture. Today watchlist, holdings, trade log, and user preferences live in browser localStorage. To become a real product these need to sync across devices, survive browser cache clears, and support audit/compliance later. How would you migrate this without breaking the current single-user experience? Specifically: would you push the localStorage data to dedicated Postgres tables behind new REST endpoints, or use a more event-sourced approach? Show me you have done this kind of migration before, ideally with a link to a similar project.
2. Broker integration & regulatory boundaries. The trade builder currently outputs a ThinkorSwim order string for the user to paste manually. The obvious next step is direct broker integration (Alpaca, Tradier, IBKR, Schwab API). What does production-grade order routing look like in your experience — idempotency keys, partial fills, websocket reconnect, paper-vs-live toggling? Equally important: how do you handle the regulatory gray zone where "signals" and "recommendations" can be construed as investment advice in the US? Have you worked under SEC/FINRA constraints before, and what disclaimer + UX patterns did you use?
3. Monetization & growth-readiness. Right now this is a tool, not a product. To charge users $20-50/month I need: tiered plans, Stripe billing, usage metering on AI Assist tokens (Claude is not free), trial flow, and analytics that tell me which features drive retention. What is your concrete recommendation for sequencing these — what ships first, second, third — and which third-party services (Stripe, PostHog, Sentry, others?) would you reach for? Bonus: have you ever built the unit economics model that decides whether $30/month is sustainable given Polygon + Anthropic data costs?
Clean, opinionated code; engaged daily collaboration; honest feedback; full credit and reference for your portfolio. Code review from someone who reads diffs carefully. No dead-air weeks.
If this resonates with you, please apply with answers to the three questions above plus 2-3 links to comparable work. I will respond to every serious proposal within 48 hours.