Senior Machine Learning Engineer

Fusemachines•New York, NY

6d•Remote

About The Position

We’re hiring a Senior Machine Learning Engineer to architect, build, and deploy high-performance machine learning systems that power technology stack. You will work across the entire ML lifecycle—from processing massive volumes of data to developing and deploying low-latency models. You must possess a strong hybrid skill set: deep expertise in applied machine learning combined with production-grade software engineering skills. You will not just build models in notebooks; you will write scalable, production-ready code, design real-time inference APIs, and ensure your systems meet strict latency and high-throughput requirements.

Requirements

5–8+ years of experience as a Machine Learning Engineer or Software Engineer focusing on ML systems, ideally within Ad Tech, MarTech, or high-scale recommendation systems.
Production Engineering Skills: Strong software engineering fundamentals (OOP, data structures, algorithm design). Expert-level Python and strong proficiency in a compiled or high-performance language (e.g., C++, Java, Scala, Go, or Rust).
ML Systems & Serving: Deep experience deploying machine learning models into highly concurrent, low-latency production environments (APIs, microservices, Triton Inference Server, custom containers).
Distributed Computing: Hands-on experience with big data processing (Apache Spark, Kafka, Flink) and complex SQL queries.
Core ML & Deep Learning: Proven track record of shipping both tree-based models and neural networks (PyTorch/TensorFlow) to production.
Statistics & Experimentation: Solid grasp of statistics, hypothesis testing, and rigorous A/B experiment design.

Nice To Haves

Agentic / GenAI Development: Experience designing agentic workflows or utilizing LLMs to automate ad creative generation, campaign copilot tools, or internal ML development workflows (AI-assisted IDEs, code agents).

Responsibilities

Scale Data Engineering & Feature Pipelines: Process and extract features from massive, highly sparse datasets (terabytes/petabytes of bidstream and user event data) using SQL, Python, and distributed computing frameworks (e.g., Spark, Ray).
Architect offline and online feature pipelines. Manage real-time feature computation and low-latency feature stores ensuring zero online/offline skew.
Perform rigorous missingness analysis, leakage checks, and handle high-cardinality categorical variables safely.
Core ML & Deep Learning Development: Train, tune, and scale supervised learning models, utilizing advanced gradient boosting (XGBoost, LightGBM, CatBoost) and Factorization Machines.
Design and implement Deep Learning architectures for structured/recommendation data using PyTorch or TensorFlow.
Apply rigorous tabular modeling practices: meticulous leakage prevention, class imbalance strategies, and robust cross-validation on time-split data.
Productionization, MLOps, & System Engineering: Write clean, object-oriented, and modular production code. Transition models from Python research environments to high-performance serving environments (packaging with ONNX, TensorRT, etc).
Design and maintain robust MLOps pipelines: automated model retraining, versioning, shadow deployments, and CI/CD for machine learning.
Monitor production models for data drift, concept drift, and performance degradation in real-time, implementing automated alerting and fallback mechanisms.
Evaluation & Experimentation: Design rigorous A/B and multivariate tests to measure the true business incrementality of ML models.
Choose appropriate offline metrics (PR-AUC, normalized Entropy/LogLoss, Calibration, Lift) and bridge them to online business KPIs.