univariate and multivariate time series forecasting models
Machine Learning
Job requirements
Key Responsibilities
Partner with product and business teams to define problems and translate them into data-driven solutions.
Conduct exploratory data analysis (EDA) and extract actionable insights from structured and unstructured datasets.
Develop, validate, and iterate on predictive models using techniques in supervised, unsupervised, and/or time series learning.
Communicate modeling outcomes through clear visualizations and presentations to both technical and non-technical stakeholders.
Build and maintain robust pipelines for model training, evaluation, and inference.
Deploy machine learning models into production with attention to scalability, performance, and observability.
Monitor model drift and performance over time and develop retraining and versioning strategies.
Collaborate with software and data engineering teams to integrate ML solutions into end-user applications and internal systems.
Qualifications Required:
Extensive hands-on experience in univariate and multivariate time series forecasting models, ideally experience with models such as LGBM, Prophet, or similar.
Deployment experience, including taking forecasting models into production.
Proper vetting prior to scheduling interviews with our team to ensure alignment with these expectations.
Master’s plus degree in Computer Science, Statistics, Applied Mathematics, or a related field.
5+ years of experience in data science and machine learning, with a proven track record of delivering models to production.
Proficiency in Python and ML libraries such as scikit-learn, XGBoost, LightGBM, PyTorch, or TensorFlow.
Strong understanding of statistical modeling, machine learning algorithms, and experiment design.
Solid experience with SQL and data manipulation tools (e.g., Pandas, Spark, or Dask).
Experience deploying models using APIs (Flask, FastAPI), Docker, and orchestration tools (e.g., Airflow, Kubeflow, MLflow).
Hands-on experience with cloud platforms (AWS, GCP, or Azure) and model serving tools.
Excellent problem-solving and communication skills; able to explain complex concepts clearly and effectively.
Preferred:
Experience with time series forecasting, causal inference, recommendation systems, or NLP.
Familiarity with data versioning and reproducibility tools (e.g., DVC, Weights & Biases).
Exposure to feature stores, streaming data (e.g., Kafka), or real-time ML systems.
Background in MLOps and experience building generalizable ML frameworks or platforms.
Here is some additional context that we have put together regarding what we are looking for: Core
Technical Skills ML Engineer Preferred: Ideally, the candidate should be an ML Engineer, though seasoned Data Scientists with relevant experience are suitable.
Python & SQL: Strong coding and data manipulation skills.
Time-Series Forecasting: Experience with LGBM (LightGBM) and Darts library. MLOps Expertise Preferred: Hands-on experience with Astronomer, Airflow, and DAG creation.
Capable of building wrappers and scalable pipelines.
This skill is highly valuable, but not a deal breaker. Cloud Platforms: Proficient in AWS, with exposure to GCP preferred.
Debugging & Troubleshooting: Skilled in investigating and resolving issues in Python experiments and executions.
GitHub Proficiency: Comfortable working in repositories with many contributors, managing branches, pull requests, and code reviews. Collaboration & Work Style
Self-Starter: Able to work independently and proactively contribute ideas.
Team-Oriented: Willing to support Roman and Calvin while offering directional guidance on model enhancements.
Fast Learner: Quick to adapt to new tools, workflows, and business contexts to rapidly onboard into the project.
Domain Expertise Sales Forecasting: Proven experience in building and refining forecasting models. Understanding of business KPIs and translating insights into action.
