Founding Full-Stack Engineer

Há 1 dia

Mexico City, Sao Paulo, Brasil Rad Hires Tempo integral

Our client is building the AI intelligence layer for commercial real estate inside Excel. Think Cursor or Endex, but purpose-built for institutional CRE acquisitions. An AI analyst that lives inside the analyst's existing workbook, understands CRE-specific models, and can do everything from parsing an Offering Memorandum to building a DCF to running market research, all with full provenance and human-in-the-loop control on every action.

We're looking for a Founding Full-Stack Engineer (CTO-track) to co-own the platform. This is not a greenfield build. There is a substantial, well-architected codebase across Python, TypeScript, React, and PostgreSQL. You'll read and own existing code, think architecturally, ship end-to-end, and learn the CRE domain.

WHAT YOU'LL BUILD

Conversational modeling: natural language to deterministic cell writes in the user's proprietary Excel template. Sensitivity tables, what-if scenarios, and formula generation.
Document intelligence: extract structured data from OMs, P&Ls, and Rent Rolls with per-field confidence and bounding-box citations. Every broker formats differently.
Template mapping: AI reads the user's workbook structure and maps extracted fields to the right cells. Not a generic template, their 12-tab institutional model, their cell layout.
Multi-model orchestration: GPT-5.1 planning loop with sub-agent delegation, routing tasks to the right model (OpenAI, Anthropic, Gemini, Perplexity, Mistral) by cost, quality, and latency.
Report generation: LOIs, investment memos, and due diligence checklists generated from the populated model.
Provenance on everything: every cell traces to source document, page, and bounding box. Can't cite, flagged as an assumption.

THE STACK

Frontend: React 19, TypeScript, Office.js, TanStack Query 5, Zustand 5, Tailwind, shadcn/ui, Webpack 5
Gateway: Next.js (App Router), Auth0 v4, Drizzle ORM, Azure Blob Storage
Orchestrator: FastAPI (async Python), SSE streaming, coordinator + 8 sub-agents
AI / LLM: GPT-5.1 / 4.1, Claude, Gemini 2.0 Flash, Perplexity Sonar Pro, Mistral, pgvector RAG
Parsers: Stateless pipelines, Azure Doc Intelligence + Mistral OCR
Database: PostgreSQL with RLS, async Python + Drizzle ORM
Infra: Azure (Container Apps, Service Bus, Blob), Redis, Docker Compose, Turborepo
Testing: Playwright E2E, TypeScript strict across all packages

WHAT WE'RE LOOKING FOR

Must-Have

TypeScript + React: complex React 19 add-in with Office.js constraints and Next.js gateway. Daily work across both.
Python (strong): orchestrator, parsers, LLM integration. Thousands of lines of async Python to read, own, and extend.
PostgreSQL: large schema with RLS policies, migrations, schema design decisions.
API design: 3-tier boundary enforcement, REST, SSE streaming, Service Bus. You care about where truth lives.
LLM integration: tool-calling, prompt engineering, RAG, context window management in production.
Self-starter: small team, big surface area. You read code and make decisions without waiting for direction.

Strong-to-Have

RAG + vector search: embedding pipelines, hybrid retrieval, pgvector.
Evals + labeling: parser quality harnesses, LLM output scoring, feedback loops.
Office.js experience (unusual runtime, willingness to learn is fine).
Azure services: Blob, Doc Intelligence, Service Bus, OpenAI on Azure.
Financial domain knowledge: CRE underwriting, P&L analysis, rent rolls, learnable on the job.

YOUR FIRST 30 DAYS

Week 1: Local dev setup. Run the full stack. Execute every user workflow end-to-end. Trace a chat message across all three tiers.
Week 2: Trace a parser run end-to-end. Read the orchestrator and agent architecture. Understand model routing and sub-agent delegation.
Week 3: Scoped improvement, parser quality, normalization, or a new capability. Fix one documented boundary violation.
Week 4: Own a feature end-to-end. Drive an integration task (wire a new model, implement a detection pass, ship a UI improvement).

WHY THIS ROLE

CTO-track: founding engineering role with a direct path to CTO as the team scales.
Real architecture: substantial, well-architected codebase with documented tech debt and clear priorities.
Domain with teeth: institutional CRE is a $20T+ market where same-day OM-to-model

Receba alertas de vagas semelhantes

Receba vagas de emprego de Founding Full-Stack Engineer

Founding Full-Stack Engineer

Receba alertas de vagas semelhantes

Vagas similares em Mexico City

Vagas similares

Américas

Europa

Ásia / Oceania

África

Founding Full-Stack Engineer

Receba alertas de vagas semelhantes

Vagas similares em Mexico City

Vagas similares