Data Scientist

Sigma Software; United States; Remote; USD 140,000 - 190,000; Full time

Job summary

Sigma Software is seeking a Senior Machine Learning Engineer to build production ML for a real-time ad exchange, handling hundreds of millions of requests daily. You will contribute to bid optimization, calibration, and counterfactual evaluation within a scalable ML platform.

You will collaborate with data scientists and engineers to design models for predictive decisioning, audience targeting, and multi-objective optimization, advancing end-to-end ML workflows from experimentation to deployment.

Qualifications

• 5+ years of ML/DS experience with production-grade models.
• Proficient Python with numpy, pandas, and scikit-learn.
• Strong SQL skills and large-scale data handling.
• Hands-on with XGBoost, LightGBM, or CatBoost.
• Experience with A/B testing and offline/online evaluation.
• Knowledge of calibration methods and SHAP/interpretability.
• English at Upper-Intermediate level or higher.
• Experience in AdTech modeling including CTR/CVR and RTB systems.

Responsibilities

• Build censored bid-landscape models to estimate clearing-price distributions from partially observed auction data.
• Develop real-time win probability estimation models responsive to bid pricing dynamics.
• Design and implement hierarchical lift estimation models with confidence-bound-based selection strategies.
• Build conversion propensity models using sparse, delayed, and aggregate-only labels.
• Develop look-alike audience modeling approaches using positive-unlabeled learning and embedding-based nearest-neighbor techniques.
• Implement advertiser-level calibration strategies while monitoring ranking and calibration quality.
• Design robust offline evaluation frameworks using inverse-propensity scoring, doubly-robust estimators, and importance reweighting.
• Define exploration strategies and propensity logging approaches to ensure reliable downstream correction and evaluation.
• Develop constrained optimization mechanisms for campaign objectives, pricing constraints, and volume targeting.
• Contribute to data diagnostics, capability assessments, and evidence-based model recommendations.
• Collaborate with Customer data science team during post-launch tuning and validation.

Skills

Python

SQL

Spark/PySpark

XGBoost/LightGBM/CatBoost

A/B testing

SHAP interpretation

Feature engineering

Statistics

Vertex AI

English communication

Tools

Vertex AI

Job description

Описание

Join Sigma Software to help build advanced machine learning solutions for one of the large-scale players in the programmatic advertising ecosystem. We are looking for a Senior Machine Learning Engineer with strong production ML expertise and deep interest in real-time optimization systems, large-scale behavioral data, and AdTech challenges. In this role, you will work with a dedicated Sigma Software team on a predictive modeling platform integrated with a live ad exchange processing hundreds of millions of auction requests daily. You will contribute to sophisticated ML solutions involving bid optimization, calibration, counterfactual evaluation, and constrained decision-making systems. We as a company offer the opportunity to work on technically challenging products, collaborate with experienced engineers and data scientists, and make a direct impact on large-scale production systems.

CUSTOMER

Our Customer is a technology company operating supply-side infrastructure within the programmatic advertising ecosystem. The company manages a high-scale ad exchange platform and is investing in predictive decisioning capabilities to improve advertising performance, audience targeting, and campaign optimization through advanced machine learning technologies.

PROJECT

The project focuses on building a predictive modeling and optimization platform on top of a live ad exchange environment. The platform evaluates and filters advertising supply in real time, predicts high-performing audience contexts, builds look-alike audiences from small seed datasets, and optimizes campaign performance across multiple business objectives and operational constraints. The team works on complex machine learning challenges including censored bid-landscape modeling, sparse and delayed conversion attribution, calibration systems, counterfactual evaluation, and constrained optimization models. The solution is designed for large-scale production use and close collaboration with the Customer’s internal data science organization.

Задачи

• Build and improve censored bid-landscape models to estimate clearing-price distributions from partially observed auction data
• Develop real-time win probability estimation models responsive to bid pricing dynamics
• Design and implement hierarchical lift estimation models with confidence-bound-based selection strategies
• Build conversion propensity models using sparse, delayed, and aggregate-only labels
• Develop look-alike audience modeling approaches using positive-unlabeled learning and embedding-based nearest-neighbor techniques
• Implement advertiser-level calibration strategies while independently monitoring ranking and calibration quality
• Design robust offline evaluation frameworks using inverse-propensity scoring, doubly-robust estimators, and importance reweighting
• Define exploration strategies and propensity logging approaches to ensure reliable downstream correction and evaluation
• Develop constrained optimization mechanisms for campaign objectives, pricing constraints, and volume targeting
• Contribute to data diagnostics, capability assessments, and evidence-based model recommendations
• Collaborate with the Customer team during post-launch tuning and performance validation cycles
• Prepare technical documentation and knowledge transfer materials for the Customer’s internal data science team
• Participate in architecture discussions and contribute to scalable ML platform design decisions

Требования

• 5+ years of experience in Machine Learning or Data Science with production-grade models measured against business KPIs
• Strong Python skills including numpy, pandas, and scikit-learn
• Strong SQL skills and experience working with large-scale datasets
• Deep practical experience with XGBoost, LightGBM, or CatBoost
• Strong understanding of regularization, calibration methods, and categorical feature handling
• Strong knowledge of probability, statistics, confidence intervals, and statistical power analysis
• Experience with feature engineering for structured and behavioral datasets
• Hands-on experience with Spark or PySpark
• Practical knowledge of experimentation frameworks and A/B testing methodologies
• Experience with advanced validation approaches including temporal splits, leakage detection, drift analysis, and slice-based metrics
• Understanding of explainability techniques such as SHAP and permutation importance
• Upper-Intermediate English level or higher
• Experience in AdTech modeling including CTR/CVR prediction, bid-landscape modeling, audience segmentation, and RTB mechanics
• Experience working with sparse, delayed, or censored labels
• Knowledge of attribution modeling, survival analysis, and positive-unlabeled learning
• Practical experience with counterfactual and off-policy evaluation techniques
• Understanding of calibration methods including isotonic regression and Platt scaling
• Experience with hierarchical, empirical-Bayes, or partial-pooling models
• Knowledge of constrained or multi-objective optimization approaches
• Experience with uplift modeling and causal inference methods
• Experience with Vertex AI or similar managed ML training environments
• Publications, competitive modeling achievements, or open-source contributions related to Machine Learning or AdTech

PERSONAL PROFILE

• Strong analytical and problem-solving skills
• Ability to work effectively in a highly data-driven environment
• Strong communication and stakeholder management abilities
• Ability to explain complex modeling decisions to technical and non-technical audiences
• Proactive mindset with strong ownership mentality
• Attention to detail and scientific rigor in experimentation and evaluation