AI / ML Engineer

You build the core of the product. Real-time voice analysis, in-call coaching triggers, post-call scoring. You will work directly with the founding team on the models that power every eNZeTi session. This is not a research position. Every model you ship runs in production during live law firm intake calls, coaching coordinators in real time. Sub-200ms latency is not a stretch goal. It is a requirement., * Train and fine-tune transformer models on legal intake conversation data

• Engineer sub-200ms inference pipelines for in-ear real-time coaching delivery
• Build post-call scoring models that grade coordinator performance across 10+ dimensions
• Own model evaluation, quality metrics, and continuous improvement loops
• Collaborate with backend engineering on streaming audio architecture
• Define the long-term ML architecture as data volume and customer base scale
• Mentor future ML team hires as the engineering org grows, * Real call data from active law firm customers - labeled, structured, growing daily
• A founding team that understands AI deeply and will not slow you down
• Greenfield ML infrastructure - no legacy debt, your architectural decisions stick
• A clear roadmap from coaching prompts to predictive case outcome scoring

Why This Role Matters

94% of law firm intake calls go completely unreviewed. Coordinators get no feedback, no coaching, no real-time support on calls worth $5,000 to $50,000 in case value. What you build changes that. Every model improvement is a law firm that closes more cases, a coordinator who gets better at their job, and a client who actually gets the legal help they called about.

• 3+ years shipping production ML systems - not notebooks, not prototypes
• Real-time audio, speech, or conversational AI experience (non-negotiable)
• Python, PyTorch or TensorFlow, AWS or equivalent cloud infrastructure
• Strong understanding of intent classification, NER, and dialogue systems
• Experience optimizing model inference for latency-sensitive production environments
• Ability to work directly with founders and ship without hand-holding