Title: Senior Full-Stack Developer / LLM Consultant (Contract: 6 Weeks)
Project: Relei — AI-Powered Proposal Direction MVP
The Role
We are seeking an experienced, pragmatic full-stack contractor to rapidly rebuild and deliver a Minimum Viable Product (MVP) for our B2B consulting platform, Relei. This 6-week engagement requires you to own the entire technical stack, transforming an existing code foundation and a detailed specification into a stable, pilot-ready application. Success hinges on exceptional LLM prompt engineering and a commitment to MVP principles (build simple, avoid unnecessary complexity).
Key Responsibilities:
- Full-Stack Development: Build the backend (Python/Node.js) and frontend (React/SPA) from scratch/existing partial codebases within a 6-week deadline.
- LLM Integration & Prompt Engineering: Integrate with commercial LLMs (e.g., OpenAI, Anthropic) and develop high-quality, structured prompts to generate a Structured Brief, Point of View (PoV), and Proposal Outline.
- Data Handling: Implement file ingestion for DOCX/PDF/CSV and handle multi-agency configuration via simple JSON files.
- Multi-Config Architecture: Design the backend to load and utilize different agency configuration files per request (B++ Multi-Config Mode) without requiring a database or user login.
- Core Feature Delivery: Implement the PoV generation and the minimal PPTX slide export focusing on "Understanding of the Ask" and "PoV" slides.
Required Technical Skills:
- Full-Stack: Expert proficiency in Python (e.g., Flask/FastAPI) or Node.js (Express) AND a modern frontend framework (React preferred).
- LLM/AI: Proven experience in using commercial LLM APIs (OpenAI, Anthropic, etc.), prompt engineering, and implementing guardrails.
- Data & File Handling: Experience with text extraction from common formats (DOCX/PDF) and CSV parsing.
- Utilities: Familiarity with a PPTX generation library (e.g., python-pptx or PptxGenJS).
- Ways of Working: Experience with Agile/MVP development, frequent check-ins, and Git/GitHub.
Non-Goals (Avoid):
- No User Authentication, Database, or true SaaS Multi-Tenancy.
- No Complex Multi-Agent Systems or Deep Vector DB/Knowledge Graph ingestion.