Senior Full-Stack Engineer (AI-Experience)

Remote — Full-time Contractor (Pathway to Permanent Employment & Potential Relocation to Australia)

About Evatt AI

Evatt AI is an emerging AI platform for lawyers and legal professionals. Our goal is to make advanced legal reasoning and document understanding accessible through natural language.

Our stack integrates Next.js, Python FastAPI, vector search, and LLM-based retrieval-augmented generation (RAG) to deliver high-quality, legally grounded insights.

We are entering a new phase — expanding beyond a chat-based interface toward a legal casebase system similar to JADE.io or AustLII, where users can perform natural language search across case law, legislation, and knowledge bases.

This is a high-autonomy role. You will work directly with the founder, take ownership of major milestones, and lead the technical direction of the product end-to-end.

Responsibilities

• Take full technical ownership of Evatt AI’s codebase (Next.js + FastAPI + Dockerized microservices).

• Lead the development of new core modules, including:

• A searchable legal casebase powered by LLMs and vector databases (RAG pipeline).

• Enhanced AI streaming, query generation, and retrieval architecture.

• Frontend refactor to modular React components for scalability.

• A modern document ingestion pipeline for structured and unstructured legal data.

• Manage releases, testing, deployment, and production stability across staging and production environments.

• Work directly with the founder to define and deliver quarterly technical milestones.

• Write clean, well-documented, production-grade code and automate CI/CD workflows.

Required Technical Skills

Core Stack (Current Evatt AI Architecture):

• Frontend: Next.js 15, React 19, Tailwind CSS, Material UI (MUI)
• Backend / API Gateway: Node.js, TypeScript, Drizzle ORM, Zustand (state management)
• AI Services: Python 3.11+, FastAPI, Pydantic, Starlette, Uvicorn
• Databases: PostgreSQL (Railway), MySQL (local), Drizzle ORM
• Vector Database: Pinecone (experience with Qdrant or Milvus is a plus)
• LLM Providers: OpenRouter, OpenAI, Google Gemini, Anthropic Claude
• Embeddings & NLP: sentence-transformers, Hugging Face, scikit-learn, PyTorch
• Containerization: Docker, Docker Compose (local dev)
• Cloud Deployment: Railway or equivalent PaaS
• Auth & Payments: Google OAuth 2.0, Better Auth, Stripe (webhooks, subscriptions)
• Email & Communication: SendGrid transactional email, DKIM/SPF setup

Future Stack (Desired Familiarity):

• Building vector-based legal knowledge systems (indexing, semantic search, chunking)
• React component design systems (refactoring from monolithic Next.js areas)
• Legal text analytics / NLP pipelines for case law and legislation
• Elasticsearch / Qdrant / Weaviate integration for advanced retrieval
• Open-source RAG frameworks (LangChain, LlamaIndex) or custom RAG orchestration
• Software architecture, prompt engineering, and model orchestration
• CI/CD pipelines (GitHub Actions, Railway deploy hooks)
• Performance, latency and scalability optimization

Soft Skills & Work Style

• Highly autonomous; able to operate without day-to-day supervision - well suited to former freelance developer or solo founder
• Comfortable working directly with a founder and delivering against milestones
• Strong written and verbal communication
• Ownership-driven; cares about reliability, UX, and long-term maintainability

Technical Interview Project

Goal: show that you can design and implement a small but realistic AI-powered legal information system.

Example challenge – “Mini Legal Casebase Search Engine”:

Build a prototype of a web-based tool that:

1. Accepts upload of legal case summaries or judgments (PDF or text).
2. Converts and embeds these documents into a vector database (Pinecone, Qdrant, or similar).
3. Supports natural language search queries such as “breach of contract in retail” and returns semantically relevant cases.
4. Displays results ranked by relevance, with extracted snippets or highlights for context.

Evaluation criteria:

• Clear, sensible architecture (frontend/backend separation, RAG flow is obvious)
• Clean, modular, documented code
• Quality/relevance of retrieval
• Bonus: simple UI with streaming AI-generated summaries

Role Type & Benefits

• Engagement: Full-time contractor (40 hrs/week)
• Transition: Potential to convert to full-time employment after 3–6 months, based on performance
• Compensation: Competitive and scalable with experience; paid monthly
• Growth path: Long-term contributors may be offered the opportunity to relocate to Australia
• Remote policy: Must be based within ±3 hours of GMT+8 (India, China, Singapore, Malaysia, Western Australia)

How to Apply

Send an email to ashley@evatt.ai with:

• Subject: “Evatt AI – Full-Stack AI Engineer Application”
• A short cover letter outlining your experience with AI systems or legal-tech products
• A GitHub & portfolio link with previous work (especially AI or RAG-related projects)
• (Optional) A short proposal outlining how you would approach building a “legal casebase search engine” similar to JADE.io / AustLII (You'll be required to build a prototype in the technical interview - so this is strongly recommended)