Senior Machine Learning Engineer (US based)
About the client
Founded in 2024, this rapidly scaling AI startup is building next-generation infrastructure for one of the most operationally complex industries in the U.S.: transportation and logistics. Their mission is to bring modern, ethical, and accessible AI technology to an industry that still relies heavily on manual workflows, fragmented systems, and regulatory guesswork.
Product: A specialized AI platform designed to automate compliance, safety management, and administrative processes for carriers. From regulatory questions to documentation, audits, driver qualification, and operational support, the platform streamlines the back-office burden so carriers can stay compliant, reduce risk, and focus on running their fleets.
Traction: Adopted by carriers seeking to modernize operations and eliminate inefficiencies, the platform has quickly gained attention for delivering tangible time-savings, clearer regulatory guidance, and AI-powered decision support tailored to real-world transportation workflows.
Why it matters: Compliance in the transportation sector is costly, time-consuming, and high-stakes. By automating regulatory complexity and reducing administrative load, this company is helping carriers operate more efficiently, improve safety outcomes, and build more resilient supply chains. Their work is accelerating the modernization of a critical industry — one that moves the world.
About the Role
We’re seeking a Senior Machine Learning Engineer to build and scale production-grade AI systems that power next-generation products. This is an applied, hands-on engineering role focused on designing, evaluating, and deploying LLM-centric pipelines with strong RAG and orchestration components.
You’ll work across multiple modalities—text, images, and audio—owning the full ML lifecycle from data ingestion to inference and evaluation. Your mission will be to create reliable, performant, and safe agentic systems that can plan, reason, and act across complex real-world workflows.
What We’re Looking For
- 7–10+ years of combined ML and software engineering experience.
- 5+ years of professional Python experience.
- 3+ years building and deploying LLM, RAG, or NLP systems in production environments.
- Python (expert): Production-level code quality, testing, and CI/CD discipline.
- LLMs & NLP: Proven experience with frontier model APIs (e.g., GPT, Claude) in production environments.
- RAG & Search: Proficiency in vector and hybrid retrieval (BM25 + dense), reranking, and chunking strategies.
- Modeling Toolkit: Hands-on experience with PyTorch and the Hugging Face ecosystem (Transformers, Datasets, PEFT/LoRA).
- Data Engineering: Solid understanding of SQL (PostgreSQL), efficient processing (Pandas, Arrow), and Parquet workflows.
- Evals & Observability: Skilled in building task metrics, dashboards, and drift/cost/latency alerting systems.
- APIs & Integration: Capable of exposing ML models as services (HTTP/gRPC) with well-defined structured outputs.
Bonus Skills
- Experience building agentic or multi-agent systems, leveraging frameworks such as LangGraph, MCP, AutoGen, or CrewAI.
- Practical knowledge of fine-tuning and optimization (LoRA, QLoRA, quantization) with an eye for cost-performance trade-offs.
- Exposure to multimodal pipelines including layout-aware OCR, image understanding, and ASR.
- Familiarity with guardrails, safety testing, and red-teaming for LLM systems.
- Working knowledge of React/TypeScript to prototype and debug end-to-end features.
- Domain experience in logistics, supply chain, TMS, or EDI systems is a plus.
- TypeScript collaboration skills within shared SDK or schema-based systems.
What You’ll Do
- Translate product requirements into robust ML pipelines (RAG, structured outputs, light fine-tuning) and deliver them to production.
- Design, implement, and operate agentic workflows with planning, tool selection, memory, retries/fallbacks, and human-in-the-loop review.
- Build and maintain data pipelines for ingestion, cleaning, task labeling, and synthetic data generation across text, document images (OCR/layout), and audio (ASR).
- Develop retrieval systems (chunking, hybrid search, reranking) with continuously updated indices for high recall and low latency.
- Implement evaluation harnesses—task metrics, regression tests, red-team suites—and drive improvement through A/B testing and user feedback.
- Optimize inference quality, cost, and latency using caching, batching, quantization, and thoughtful model/API selection.
- Add safety guardrails (prompt-injection defenses, PII filtering, structured JSON outputs).
- Collaborate closely with product and full-stack engineers to integrate models into user-facing features.
- Document design trade-offs, contribute to internal SDKs and utilities, and mentor team members.
What We Offer
- USD 160k-220k with equity
- PTO Policy
- U.S Public Holidays
- The opportunity to design and operate real-world AI systems that push the frontier of applied LLMs.
- A collaborative team of experienced engineers who balance rigor with speed.
If this sounds like you, we’d love to hear from you!
