Who We Are
You are looking for a role that challenges you intellectually and lets you build things that matter. You have an entrepreneurial mindset, you enjoy solving complex technical problems, and you thrive in a fast-paced startup environment. Most of all, you want your work to have meaning.
We're Darrow – a fast-growing startup in the Legal Exposure Management space. We have an open, action-driven culture unlike any other. Our work starts from a simple idea: legal exposure should be visible while it's still possible to act on it. We've built a technology engine that scans massive amounts of public data to identify legal exposure and evaluate its financial value at sale, and every tool or workflow you build here moves that mission forward.
We have one advantage technology and AI can't replicate: our people. More than 100 people worldwide, in Israel, US & Europe who are smart, humble, curious people who thrive at the intersection of law and AI, and we're looking for the leader who will protect and grow that edge as we scale.
About the Role
A Senior AI/ML Engineer role on our Data team, sitting at the intersection of AI engineering, data science, and data platform. You'll build agentic, LLM-based systems that turn vast amounts of unstructured information into high-quality structured signals — powering Darrow's products and surfacing legal risk — owning each solution across the full lifecycle, from exploration to production. You'll work closely with data engineers, data scientists, product managers, and legal experts to turn ambiguous problems into scalable AI capabilities. This is a hands-on role for someone with strong engineering judgment, deep data/ML expertise, and a focus on reliability beyond the prototype stage.
What You'll Do
Design, build, deploy, and operate production AI systems: agentic, LLM-based extraction workflows and batch/online inference pipelines and owning them from exploration through deployment, monitoring, and iteration.
Define quality metrics and build evaluation datasets, testing processes, and feedback loops for ML and LLM systems; monitor accuracy, latency, cost, and data quality, and investigate and fix failures.
Explore large, complex datasets to surface useful features and patterns, and translate findings into production-grade implementations.
Partner with data engineers on ingestion, transformation, orchestration, and storage to deliver dependable data products.
Use AI-assisted coding tools (e.g., Claude) effectively while maintaining strong technical judgment and ownership of your systems.
Help shape engineering standards for production AI: testing, observability, reproducibility, data lineage, versioning, and safe deployment.
Responsibilities
Requirements
What You Bring
6+ years in ML engineering, AI engineering, data science, or backend engineering for data-intensive systems, with hands-on experience taking systems to production (deployment, monitoring, evaluation, versioning, improvement).
Strong Python skills and data orientation — SQL, NoSQL, data pipelines, and both structured and unstructured data.
Solid data science foundation: feature engineering, statistical reasoning, experiment design, error analysis, and relevant metrics (precision, recall, accuracy, coverage).
Experience with LLM-based applications: structured extraction, tool calling, agentic workflows, prompt design, embeddings, retrieval, and model evaluation.
Understanding of production AI tradeoffs, quality, latency, scalability, reliability, cost and experience integrating AI into batch, orchestration, API, or event-driven workflows.
Ability to work independently in ambiguous environments, breaking problems down and driving them to production.
Strong communication skills across engineering, data science, product, and domain-expert teams.
Nice to Have
Document intelligence, information extraction, entity resolution, classification, ranking, or large-scale enrichment pipelines.
Experience building or operating agentic systems in production.
Natural Language Processing experience.
Familiarity with orchestration/ML tooling (Airflow, Dagster, Prefect, MLflow, or similar).
Experience with data warehouses, data lakes, vector databases, or modern data-processing frameworks.
Cloud-native experience: AWS, Docker, Kubernetes, IaC, CI/CD.
Experience with human-in-the-loop review or feedback-driven AI systems.
Familiarity with legal, financial, compliance, or other data-intensive domains where quality and explainability matter.
What Success Looks Like
AI capabilities move reliably from experimentation to production, on systems engineers, scientists, and domain experts can depend on.
Agentic workflows deliver accurate, measurable structured data that strengthens Darrow's products and decision-making.
What You Bring
6+ years in ML engineering, AI engineering, data science, or backend engineering for data-intensive systems, with hands-on experience taking systems to production (deployment, monitoring, evaluation, versioning, improvement).
Strong Python skills and data orientation — SQL, NoSQL, data pipelines, and both structured and unstructured data.
Solid data science foundation: feature engineering, statistical reasoning, experiment design, error analysis, and relevant metrics (precision, recall, accuracy, coverage).
Experience with LLM-based applications: structured extraction, tool calling, agentic workflows, prompt design, embeddings, retrieval, and model evaluation.
Understanding of production AI tradeoffs, quality, latency, scalability, reliability, cost and experience integrating AI into batch, orchestration, API, or event-driven workflows.
Ability to work independently in ambiguous environments, breaking problems down and driving them to production.
Strong communication skills across engineering, data science, product, and domain-expert teams.
Nice to Have
Document intelligence, information extraction, entity resolution, classification, ranking, or large-scale enrichment pipelines.
Experience building or operating agentic systems in production.
Natural Language Processing experience.
Familiarity with orchestration/ML tooling (Airflow, Dagster, Prefect, MLflow, or similar).
Experience with data warehouses, data lakes, vector databases, or modern data-processing frameworks.
Cloud-native experience: AWS, Docker, Kubernetes, IaC, CI/CD.
Experience with human-in-the-loop review or feedback-driven AI systems.
Familiarity with legal, financial, compliance, or other data-intensive domains where quality and explainability matter.
What Success Looks Like
AI capabilities move reliably from experimentation to production, on systems engineers, scientists, and domain experts can depend on.
Agentic workflows deliver accurate, measurable structured data that strengthens Darrow's products and decision-making.
