HMH
HMH
HMH develops curriculum, assessment, and learning resources for K–12 education. Its programs span literacy, mathematics, science, social studies, and social-emotional learning, alongside professional development that helps educators strengthen instructional practice and support student learning.

Machine Learning Engineer, AWS SageMaker

Bring explainable student assessment models from development into production on AWS SageMaker. Support HMH learning products through reliable modeling, deployment, monitoring, and ongoing improvement.

Tavsif

  • Own the model lifecycle from data preparation and training through deployment, monitoring, and version updates
  • Explore and validate data before and after model fitting, documenting your findings
  • Train and validate models in Python with SageMaker training jobs and pipelines
  • Record experiments in MLflow and test models on held-out cohorts before promotion
  • Maintain clear lineage for model data, code, weights, and serving as you register, promote, and deploy versions
  • Release models through code review and continuous integration
  • Serve models through real-time endpoints and batch scoring with a feature store and Snowflake
  • Optimize model latency, cost, and capacity
  • Track prediction quality, failures, and drift using dashboards and alerts
  • Review production health monthly and lead blameless postmortems
  • Use production evidence to recommend model improvements and inform team RFCs
  • Help analytics, product, and data teams interpret and apply model outputs
  • Work with senior or staff engineers on stream-aligned teams
  • Develop models that generalize across subjects or curricula without separate code forks
  • Re-score large student populations while live services remain available
  • Operate model versions side by side and retire them safely
  • Improve training and serving efficiency while preserving model outputs

Talablar

  • Demonstrated experience delivering models or machine learning systems that others rely on
  • For IC4, typically at least three years of hands-on experience, or a doctorate paired with relevant practical experience
  • A master’s or doctoral degree in a quantitative discipline, including psychometrics, is expected
  • Write idiomatic, typed, tested Python
  • Comfort with CI/CD, GitHub Actions, and infrastructure managed with Terraform
  • Strong SQL skills
  • A grounding in statistics, including posterior and likelihood concepts and calibration plots
  • Proficiency in exploratory data analysis
  • Excellent technical writing skills
  • Working knowledge of AWS, preferably SageMaker
  • Working knowledge of an experiment tracking tool such as MLflow
  • Ability to judge when work is ready to ship and support that judgment with evidence
  • Helpful experience or knowledge includes psychometrics or educational measurement; cognitive diagnostic models or Bayesian networks; probabilistic programming with Stan or PyMC; R; Snowflake and dbt; Terraform; GitHub Actions; knowledge graphs or ontologies; and K–12 assessment data
  • Must be authorized to work in the United States; HMH participates in E-Verify

Imtiyozlar

  • Annual salary range of $110,000–$115,000
  • Commitment to equal employment opportunity
  • A work environment free of discrimination and harassment

O‘xshash ish o‘rinlari

CrowdStrike

Threat Analyst, Machine Learning Detection Efficacy

CrowdStrike

Assess malware and customer detection reports to help improve CrowdStrike’s machine learning detections. Support false-positive management and share efficacy findings with internal teams.

Ochish
PAR Technology

MLOps Engineer, Identity Resolution

PAR Technology

Bring Bridg’s identity-resolution models into production for PAR Technology’s restaurant software platform. Build AWS and Spark pipelines, CI/CD, monitoring, and agent-queryable data interfaces.

Ochish
Volga Partners

Remote Data Annotation Specialist, Object Detection — United States

Volga Partners

Annotate visual datasets for Volga Partners’ AI projects, labeling objects to support computer vision systems. Help improve object detection quality through careful, consistent work.

Ochish