Description
Summary Project Overview We’re looking for a skilled AI Engineer to build a smart, automated email processing pipeline. The system will intelligently classify, extract key information, and route incoming emails using Large Language Models (LLMs) and modern AI orchestration tools. This is initially a focused, high-impact project (MVP/prototype) with clear deliverables and strong potential for expansion into a full production system, ongoing maintenance, or additional AI workflow automations. Scope of Work (Phase 1 – $250) Develop a robust email ingestion pipeline (IMAP/Gmail API, webhooks, etc.). Use LLMs (e.g. Grok, Claude, GPT-4o, or open-source models via Ollama/LM Studio) for: Classification (e.g. support, sales, billing, partnership, spam, etc.) Information Extraction (entities, dates, amounts, names, intent, sentiment, etc.) Routing & Action Recommendations (auto-reply drafts, forward to team members, create tasks in CRM/Notion/Slack, archive, etc.) Implement a clean, modular architecture with logging, error handling, and basic dashboard/UI for monitoring. Provide well-documented code + setup instructions (Docker preferred). Deliverables Working prototype that processes sample emails accurately. Source code repository. Documentation + deployment guide. 1 round of revisions. Budget Phase 1 (MVP): $250 Potential Scale-Up (Phase 2+): $500+ total (or more) for production hardening, advanced features (multi-account support, custom fine-tuning, RAG integration, analytics dashboard, human-in-the-loop feedback loop, etc.), and ongoing support. Requirements Strong experience building LLM-powered automation workflows. Proficiency with email APIs, LangChain/LlamaIndex/Haystack (or similar), and orchestration tools (CrewAI, AutoGen, or custom agents). Experience with Python, Docker, and cloud deployment (AWS, Vercel, Railway, etc.). Ability to deliver clean, maintainable, and secure code. Bonus: Prior work with email automation, CRM integrations, or RAG systems. Timeline Target completion: 5–7 days from kickoff (flexible for the right candidate).