Senior Machine Learning Engineer

Liminal is the actionable intelligence company. We've built a proprietary Living Graph — a verified knowledge architecture that maps the real-time structure of Identity, Fraud, and Cybersecurity — and the agentic AI systems and human verification layer to make it trustworthy. Visa, Mastercard, Google, and JPMC use it to make strategic and revenue decisions. Series A, 80 people, offices in NYC, Salt Lake City, Porto, Lisbon, and Manila. The architecture works, the customers are real, and we're scaling the team.

Liminal doesn't wrap a generic LLM around public data and call it intelligence. We've built a compound AI system — Large Domain Models that combine neural pattern detection with symbolic resolution against the deterministic structure of a proprietary Living Graph mapping 2.5M+ entities across Identity, Fraud, and Cybersecurity. When our system reads a signal, it reasons against verified relationships. If the graph can't verify it, it doesn't ship.

We're looking for a Senior Machine Learning Engineer to build, deploy, and maintain the production-grade ML systems that power this architecture. You'll work across the full lifecycle — from exploratory analysis and proof of concept through production deployment and ongoing optimization. Reporting to the AI Solutions Architect and partnering closely with the Chief Product Officer and Chief Innovation Officer, you'll bridge the gap between experimental projects and the production systems that Visa, Mastercard, Google, and JPMC depend on daily.

This is applied ML at the intersection of knowledge graphs, NLP, and agentic systems — not fine-tuning a chatbot. The problems are hard, the data is proprietary, and the system actually works.

In your first 30 days, you've understood the architecture — the Living Graph, the Large Domain Models, and the production ML infrastructure — and shipped your first improvement to an existing model or pipeline.

In your first 90 days, you've owned a meaningful model improvement or new capability end-to-end, from design through production deployment, and established yourself as a reliable partner to engineering, data science, and product teams.

In your first year, you've shipped production ML systems that measurably improve intelligence quality, expanded the platform's reasoning capabilities, and helped define the engineering standards for how models are built, deployed, and maintained at Liminal.

• Build, deploy, and maintain production-grade machine learning models and systems that power verified intelligence across the Liminal platform
• Continuously improve models through fine-tuning, retraining, and incorporating new data, feedback loops, and Analyst Desk verification signals
• Develop and maintain pipelines for automated retraining, testing, and deployment — ensuring models stay sharp as the Living Graph grows and markets evolve
• Leverage cloud platforms (GCP, AWS, or similar) for model deployment, orchestration, and monitoring — building infrastructure that scales with growing data volumes and complexity
• Develop and maintain data pipelines for ingestion, preprocessing, and feature engineering that feed the Living Graph and downstream ML systems
• Conduct exploratory analysis to identify opportunities for AI innovation, prototype solutions through proof of concept, and transition successful experiments into production
• Implement monitoring for model performance, system reliability, and data integrity — diagnosing and resolving issues proactively
• Collaborate with the AI Solutions Architect, product engineering, and data science teams to ensure ML systems align with product goals and platform architecture
• Maintain clear documentation of workflows, models, and system designs

You've built and deployed ML systems in production and understand the difference between a model that works in a notebook and one that works at scale. You're proficient in Python and its ML frameworks (TensorFlow, PyTorch, scikit-learn), with hands-on experience across data pipeline development, feature engineering, and cloud infrastructure. You've worked with LLMs and NLP — including prompt engineering, fine-tuning, and retrieval-augmented generation — and you understand where these techniques work and where they break down. You're a strong problem solver who can debug complex systems under production constraints, and you communicate clearly with both technical and non-technical stakeholders.

Experience with knowledge graphs, neuro-symbolic architectures, or structured reasoning systems. Familiarity with agentic AI systems and autonomous signal processing. Background building compound AI systems where models, retrieval, verification, and human expertise work together. Experience at a high-growth or startup environment where you've shipped ML systems that directly drive product value.

The ML problems are real. This isn't classification on a benchmark dataset. You'll build models that reason across a structured knowledge graph with 2.5M+ entities, combine neural and symbolic approaches, and produce intelligence that's verified by human domain experts before it reaches enterprise customers. The technical challenges — retrieval over structured knowledge, domain-specific NLP, agentic signal processing — are genuinely hard.

The system actually works. Customer results are measurable: 9x faster strategic decisions, 30% higher competitive win rates, 70% reduction in manual research time. You're not building demos — you're building production ML that the world's largest platforms depend on.

Your work compounds. Every model you improve, every pipeline you build, every capability you ship makes the Living Graph smarter and the platform more valuable. You're building systems that get better with every signal, every analyst review, and every customer interaction.

We respect your time. Here's what to expect:

Recruiter Screen → Hiring Manager Interview → Behavioral Interview → Practical Interview → CEO Interview → Offer