AI Engineer - Legal Technology

About Us

We are a startup building AI-powered tools for legal case management — from the ground up, with no legacy AI codebase. Our platform serves law firms managing tens of thousands of active cases, and our AI team ships production services that process real legal documents, generate demand letters, run voice intake agents, and classify case files at scale.

This is not a research role. You will build, ship, and support production AI systems used by real people every day.

The Role

You will join a small, high-impact AI engineering team that owns the full lifecycle: architecture, implementation, deployment, monitoring, and support. Our stack spans 9 production services and 16+ active POCs across document intelligence, voice AI, agentic RAG, and generative content.

You will work directly with our CEO, CTO, domain experts, and QA team. You will not receive a 40-page requirements document — you will get a problem, explore it, prototype solutions, and iterate until it works. If that sounds uncomfortable, this role is not for you.

This is an on-site role. No remote applicants. This is not negotiable and will not be revisited.

What You Will Build

• · Document Intelligence — AI-powered classification, summarization, medical chronology extraction, and demand letter generation from legal case files (PDF, DOCX, images)
• · Voice Agents — Real-time conversational AI for legal intake using LiveKit, Deepgram, and OpenAI Realtime API
• · Agentic Systems — RAG pipelines with vector search (Chroma, HuggingFace embeddings), multi-step reasoning agents, and tool-calling LLMs
• · Production Services — FastAPI microservices deployed as Windows Services (PyInstaller + NSSM) and Docker containers on Linux
• · Infrastructure — CI/CD pipelines, Docker deployments, Grafana/Loki observability

Required Skills - Non-Negotiable

Python (10+ years professional experience)

• · Expert-level Python. Not "I use Python at work" — you think in Python
• · Async programming (asyncio, aiohttp, websockets)
• · Multi-threading, concurrency, process management
• · Package design, dependency management, virtual environments
• · Testing (pytest, mocking, fixtures, integration tests)
• · Code quality tooling (black, flake8, mypy, bandit)

AI/ML Engineering (4+ years professional experience)

• · LLM Integration: OpenAI Chat Completions, Assistants API, Realtime API, function calling, streaming. Just knowing the Chat API is not sufficient
• · RAG Systems: Vector databases (Chroma or equivalent), embedding models (HuggingFace/OpenAI), chunking strategies, retrieval pipelines
• · Agentic Patterns: Tool-calling agents, multi-step reasoning, agent orchestration frameworks (LangChain or equivalent)
• · Document AI: PDF/DOCX extraction, OCR pipelines, document classification, summarization with citations
• · Voice/Audio AI: Speech-to-text, text-to-speech, real-time audio streaming (LiveKit, Deepgram, or equivalent)
• · Prompt Engineering: Structured outputs, few-shot examples, system prompts for complex domain tasks
• · You must have shipped AI systems to production — not just Jupyter notebooks

Infrastructure & Operations

• · Windows Server: SQL Server (writing and optimizing T-SQL), IIS, Windows Services, Windows networking (SMB/CIFS, UNC paths, Active Directory basics)
• · Docker: Writing Dockerfiles, building images, managing containers, volume mounts, networking
• · CI/CD: Hands-on experience building and maintaining release pipelines — not just triggering them. You understand build artifacts, environment promotion, rollback strategies
• · Databases: SQL Server with pyodbc, connection management, stored procedures, schema understanding

Attitude & Work Style

• · Customer-oriented: Your job does not end when you push code. You work with QA to verify, support to troubleshoot, and escalation to resolve. You own the outcome, not just the commit
• · Proactive and organized: You flag risks before they become problems. You track your work without being asked. You follow up without reminders
• · Comfortable with ambiguity: We are a startup. Requirements are discovered, not delivered. You will explore, prototype, and iterate — not wait for a specification
• · Hands-on across the stack: You will write code, configure deployments, debug production issues, test edge cases, and sometimes do things that are "not your job." Startup means everyone gets their hands dirty
• · Direct communicator: You will interact with the CEO, CTO, domain experts, and QA. You need to explain technical decisions clearly and listen to non-technical feedback without defensiveness

Nice to Have

• · C#/.NET: Our main product is an ASP.NET WebForms application — ability to read and modify C# code is a plus
• · Linux Administration: Basic sysadmin skills — managing Docker hosts, troubleshooting networking, reading logs
• · Grafana/Loki: Experience with centralized logging and monitoring dashboards
• · Legal Domain Knowledge: Understanding of legal case management, medical records, insurance claims

What We Offer

• · Greenfield AI development — you are building from scratch, not maintaining someone else's decisions
• · Startup culture — small team, direct impact, no bureaucratic approval chains
• · Technology freedom — we try new tools, frameworks, and approaches. If it works better, we use it
• · Real problems — not toy demos. Your code processes real legal documents, talks to real clients, and saves real time for attorneys
• · Growth with the company — as the AI team grows, early members shape the architecture, standards, and culture

Interview Process

1. Application Review — Resume and pre-screen questionnaire. We look for evidence of shipped production systems, not course certificates

2. Technical Screen (30 min) — Technical conversation to validate depth of experience

3. Live Coding Session (90 min) — You will build a working solution to a real-world problem on a shared screen. No take-home assignments. No pre-prepared algorithms. We want to see how you think, debug, and code under realistic conditions. This step is mandatory and cannot be substituted

4. Architecture Discussion (60 min) — Walk through a system you built. We will ask about trade-offs, failures, and what you would do differently

5. Culture Fit (30 min) — Conversation with team members about working style, communication, and expectations

Pay: $90,000.00 - $150,000.00 per year

Benefits:

• 401(k)
• Dental insurance
• Health insurance
• Paid time off
• Vision insurance

Application Question(s):

• What is the most overrated AI framework or tool right now, and why?
• When is RAG the wrong approach?
• What is one task you would never delegate to an LLM?
• What is the hardest thing about running Python services on Windows?

Experience:

• Python: 8 years (Preferred)
• AI/ML: 4 years (Preferred)
• C#: 5 years (Required)
• ASP.NET: 5 years (Required)
• Microsoft SQL Server: 5 years (Required)

Ability to Commute:

• Melville, NY 11747 (Preferred)

Work Location: In person