Carlos Cueva@CuevaLabsDev
2
Events attended
2
Submissions made
United States
1 year of experience
About me
Hey, I’m Carlos Cueva, a full-stack engineer and solo founder from Newark, CA, who just loves shipping stuff that actually works. I build AI-powered tools and smart data platforms in my spare time. Right now I’m running three side projects I’m pretty hyped about: CodeSergeant: a local-LLM macOS menu-bar coach that keeps you in deep work (SwiftUI + Ollama) Trading Card Block: live market intelligence for trading card collectors with real-time pricing and population data AlphaSolver: browser-based Monte Carlo simulator for prop traders I usually work with Next.js, SwiftUI, Python pipelines, Supabase, and local LLMs. Whatever gets the product out the door fastest. I’m the guy who turns weekend ideas into working demos in a couple weeks instead of months. At this hackathon I’m looking to build something fun and useful, and probably learn a new trick or two. If you’re into AI tools, data dashboards, or just want to ship fast, come say hi. I’m always down to build.
🤓 Latest Submissions
Trading Card Block
Graded trading cards are a multi-billion-dollar alternative asset class with no unified market hub. Pricing, liquidity, grade spreads, and creator-driven momentum live across TCGplayer, eBay, and social platforms—updated continuously, rarely structured in one place.Trading Card Block solves this for serious operators: card dealers, grading businesses, marketplace sellers, and portfolio managers who need decisions on current reality, not last week's export.Bright Data is the live-web engine. Web Unlocker reaches TCGplayer's internal JSON APIs and eBay sold PSA comps without fragile Playwright scrapes. SERP API and Web Data datasets discover creator posts and full YouTube transcripts. Structured rows land in Supabase: liquidity scores, grade multiples, sentiment, and impact-ranked creators.Five intelligence layers sit on 1.9M+ daily price facts and precomputed analytics: price windows, raw to PSA arbitrage, 0–100 liquidity, creator sentiment, and cognee market memory. A Featherless AI agent with ten warehouse tools reasons across all layers; live refresh runs on demand. Vercel crons and optional Triggerware pulses fire creator_move, price_breakout, and grade_arbitrage alerts with actionable links
31 May 2026
ChefPro-AI
Campus and enterprise food service teams still run tastings, menu signage, counts, and closing checks across spreadsheets, photos, and paper logs. Compliance gaps are caught late, and leadership lacks a single source of truth across locations and meal periods. ChefPro is a multi-role operations platform built on Next.js 16, PostgreSQL/Prisma, and Supabase Storage. Chefs capture tastings and photos on mobile-first flows; ops manages campuses, deadlines, and analytics; compliance staff upload closing photos and temperature logs. Gemini analyzes closing images for potential issues managers must review, parses temperature logs into structured data and PDF archives, and powers a scoped Ops AI Assistant with tool access to tastings, packets, and compliance data. We target multi-location operators (corporate campuses, healthcare, universities) who need regulatory-grade audit trails without slowing kitchen speed. Differentiators: immutable field-level audit log, 3-tier RBAC, AI findings framed as “needs review” not auto-verdicts, and a modular monolith ready to scale. Stack: Next.js App Router, NextAuth, Prisma 7, Supabase, Gemini API, Recharts. Demo accounts are seeded for every role (see README).
19 May 2026
