Job Title: Full-Stack AI Application Engineer Company Name: Arbitration Sciences Limited Job Url: https://www.linkedin.com/jobs/view/4385545290/?eBP=NOT_ELIGIBLE_FOR_CHARGING&trk=flagship3_search_srp_jobs&refId=2gRpQsuCaTXmGFIuttG%2FOw%3D%3D&trackingId=AcSDzuwTClf3E269lJCJGQ%3D%3D Job Description: Full-Stack AI Application Engineer Arbitration Sciences Limited | Remote Miami, FL · 2 hours ago · 43 applicants Promoted by hirer · No response insights available yet Remote Full-time Easy Apply Save Use AI to assess how you fit Get AI-powered advice on this job and more exclusive features with Premium. Try Premium for Rs 0 Show match details Tailor my resume Help me stand out About the job Full-Stack AI Application Engineer Arbitration Sciences Limited | Remote We're building something that doesn't exist yet: an AI system that analyzes the psychological and linguistic patterns in international arbitration hearings and delivers real-time intelligence to legal teams. The science is done. The prompts are written. The methodology is validated. Now we need someone to build the actual product. What You'll Build A push-button application that takes raw hearing transcripts and runs them through a sophisticated multi-stage AI analysis pipeline—completely invisible to the user. Behind that simple interface, you'll be orchestrating: Parallel LLM scoring runs across multiple models Statistical reliability validation and consensus selection Semantic chunking with context window management Cross-day cumulative tracking of behavioral patterns Automated report and visualization generation You'll work directly with our research team (computational linguists, behavioral scientists, tribunal psychologists) to translate a detailed analytical methodology into production code. This is a build-from-scratch role—you'll own the entire technical stack. What We're Looking For Backend: Strong Python or Node.js development experience. You've built multi-step workflows, integrated LLM APIs (OpenAI, Anthropic, etc.) with proper error handling, and processed documents at scale. You're comfortable implementing statistical algorithms in code (consensus selection, reliability metrics, validation logic). Frontend: React, Vue, or similar—with a track record of building step-by-step workflow interfaces for non-technical users. Think wizard UIs, progress tracking, intermediate output review screens. AI/ML Integration: You've shipped production applications that use LLM APIs or other AI/ML services. You understand prompt engineering, token management, rate limiting, and retry logic. Document Processing: You can take messy PDFs or transcripts, extract clean text, chunk intelligently, and feed them into downstream processes. State Management: You know how to persist application state across sessions—this system tracks cumulative analysis across multiple hearing days. Nice to Have: Legal tech experience, NLP/behavioral science background, DevOps skills, database architecture experience. Why This Role Is Different You're not implementing someone else's architecture. You're not maintaining legacy code. You're not on a 12-person engineering team. You're the engineer building a first-of-its-kind AI product from the ground up, working directly with domain experts who've spent years developing the methodology. If you want ownership, technical autonomy, and the chance to build something genuinely novel—this is it. To Apply Send your CV and a brief note about relevant experience to info@arbitrationsciences.com In your note, tell us about: A multi-stage workflow or AI-powered application you've built (even if it's not in legal tech) Your approach to building interfaces for non-technical users Why you're interested in this particular problem