Apply on Employer Site

DAT Freight & Analytics · 1 day ago

Principal Machine Learning Engineer

Seattle, WA

Full-time

Hybrid

Lead/Staff

$245K/yr - $320K/yr

8+ years exp

DAT Freight & Analytics is an award-winning employer of choice and a next-generation SaaS technology company in transportation supply chain logistics. They are seeking a Principal Machine Learning Engineer to scale and evolve their critical Data and ML Platform capabilities, focusing on building foundational infrastructure for ML and AI systems.

LogisticsTransportation

Responsibilities

Deliver lower-latency data to models, unlocking online learning, adaptive policies, and improved real-time decision-making for Convoy’s auction mechanism, fraud detection apparatus, and carrier engagement campaigns

Evolve our ML platform to support generative AI, including orchestration, retrieval, standardized service patterns, and scalable model serving needed for foundational model applications in document digitization and voice-based features

Experiment faster and safer, through robust causal inference tooling, richer randomized experimentation, and reliable evaluation infrastructure to help us learn more about the unique spatio-temporal dynamics of a Trucking marketplace

Drive the evolution of Convoy’s experimentation and model-evaluation foundations

Enable rigorous causal measurement, reliable online experimentation, scalable model iteration, and adaptive learning systems that continuously improve marketplace and policy decisions

Evolve Convoy’s experimentation stack (TestDrive): Add richer randomized experiments, causal inference tooling, exposure/assignment logging, and metric pipelines; evaluate third-party solutions where beneficial

Enable adaptive learning approaches (RL, contextual bandits, online learning) for dynamic marketplace and policy decisions (e.g., inferring the best timing, cohort, or communication channel to maximize carrier engagement)

Harden our evaluation infrastructure, including offline/online pipelines, drift detection mechanisms, and structured feedback loops that ensure reliable model behavior over time

Implement orchestration layers that combine inference, retrieval, business logic, guardrails, and human-in-loop flows into reliable, auditable multi-step AI agents

Iterate on and expand Convoy Platform’s low-latency Feature Store and real-time streaming platform (on RisingWave) to deliver signals such as app analytics, carrier behavior, and digital fingerprints to support marketplace optimization, fraud detection, and other decision systems

Ensure unified online/offline semantics to improve online decision-making, support real-time optimization, and enable future reinforcement-learning and online-learning workflows

Build high-throughput streaming pipelines for carrier engagement, risk indicators, and fraud signals that power sub-minute marketplace and policy decisions

Develop platform-level trucking knowledge systems, including RAG indexes, domain adapters, structured benchmarks, and retrieval strategies that ground AI systems in operational realities

Design and scale the platform ecosystem, leveraging Kafka, Snowflake, Kubernetes, and modern data formats (Avro, JSON, Iceberg), and use Python/Go to build the “connective tissue” that ensures platform reliability and scale

Build low-latency, production-grade Python services and contribute to TypeScript/Node where needed (e.g., emitting high-quality data signals, wiring model calls into product flows, enabling experimentation and feature-flag pathways)

Partner with scientists to define durable service patterns (API design, serving workflows, monitoring) and uplift the platform that enables fast, safe iteration on ML-backed services

Mature platform infrastructure, including Terraform/IaC, CI/CD, observability, logging/tracing, incident readiness, and cost/performance optimization

Improve SQL/dbt workflows and batch/streaming pipelines to increase reliability, correctness, and scalability

Extend model-serving infrastructure to support more advanced ML workloads (managed inference →self-hosted GPU), with standardized versioning, canary/A/B rollouts, and granular monitoring

Qualification

Machine Learning EngineeringReal-time ML PlatformsPython ProgrammingData InfrastructureKubernetesKafkaCI/CDMLOpsCollaborationMentoringProblem Solving

Required

8–12+ years of experience in ML engineering, data infrastructure, platform engineering, or closely related production engineering roles

Deep hands-on experience with real-time ML platforms, including feature stores, stream processing, low-latency data services, and online inference systems

Strong proficiency in Python, with the ability to work across non-Python stacks including TypeScript/Node, gRPC services, and Kubernetes-based microservice ecosystems

Expertise in modern data and ML infrastructure, including Kafka, Kubernetes, Postgres-like OLTP systems, cloud platforms, and production observability tooling

Experience building and operating robust data and ML pipelines (both batch and streaming), ideally in high-scale environments such as marketplaces, fraud detection systems, pricing, personalization, or real-time decision platforms

Strong DevOps and MLOps fundamentals, including CI/CD, containerization, infrastructure-as-code (Terraform/Helm), automated monitoring, and cloud cost and performance optimization

Collaborative platform mindset, with a track record of partnering with scientists and product engineers to co-design durable service patterns for model serving, deployment, monitoring, and API design, enabling fast and safe iteration on ML-backed systems

Ability to operate at Principal scope, setting technical direction, identifying and retiring platform risk, mentoring engineers, and delivering solutions whose impact scales across teams and the broader organization

Preferred

Experience building ML systems in two-sided marketplaces, financial markets, or other economically complex environments, with intuition for incentives, pricing, and market dynamics

Deep experience with data reliability and correctness at scale, including schema evolution, data quality enforcement, backfills, late data handling, and incident response for production data systems

Applied advanced ML techniques such as reinforcement learning, bandits, or optimization to unlock real-world business impact, ideally within freight, logistics, or transportation technology