Syniverse is the world’s most connected company. Whether we’re developing the technology that enables intelligent cars to safely react to traffic changes or freeing travelers to explore by keeping their devices online wherever they go, we believe in leading the world forward. Which is why we work with some of the world’s most recognized brands. Eight of the top 10 banks. Four of the top 5 global technology companies. Over 900 communications providers. And how we’re able to provide our incredible talent with an innovative culture and great benefits.
Who We're Looking For
The Machine Learning Operations Consultant enhances the reliability, quality, and automation of Syniverse's ML model development and deployment lifecycle. This role focuses on modernizing Ground Truth (GT) dataset management, strengthening model quality assurance and validation practices, and designing a fully automated model re-training pipeline. The consultant works independently and collaboratively with the Technical Curation (TC) team and other stakeholders, delivering recommendations, designs, and implementation support through workshops, analysis, and documentation. This is an engagement-based, time-and-materials advisory role rather than a traditional full-time position. -
Some of What You'll Do
Ground Truth (GT) Maintenance
Assess current S3-based GT storage and versioning practices.
Recommend and design a scalable GT management system supporting seamless, criteria-based (age, source, quality) replacement or updating of records at the record level, while preserving vital historical data needed for domain-specific classifiers.
Evaluate migration toward a more structured storage solution (e.g., database-backed repository) to enable advanced filtering, automated quality/validation checks, and access by multiple downstream services.
Provide implementation support or a prototype for automated GT refresh and validation processes tailored to the training pipelines.
Model Quality Assurance and Validation
Evaluate the current post-training assessment approach, including use of small live-data snapshots and empirical accuracy/degradation thresholds.
Design and set up an independent integration testing framework using small, highly curated test datasets targeted to specific product requirements, run alongside the latest production data snapshots.
Recommend additional quality checks, including methods to detect and quantify data distribution shifts (data drift) between historical GT blends and live production data.
Define procedures and tools for ongoing model quality monitoring post-deployment.
Re-training Automation
Design and support implementation of a fully automated monthly re-training pipeline that ingests new material from the TC team and integrates with the updated GT management system.
Recommend best practices for handling imbalanced datasets, including evaluation of privacy-aware, LLM-based synthetic data augmentation techniques where appropriate.
Select and configure orchestration tooling (e.g., MLflow, Apache Airflow, or a custom system) to establish a formal, well-documented, and reproducible re-training process that reduces reliance on manual curation or any single individual.
Integrate GT maintenance and model quality checks into the automated pipeline, and deliver comprehensive documentation, knowledge transfer, and handover.
Program and Communication
Participate in regular progress review meetings and provide periodic status updates against the agreed work plan and milestones.
Maintain written documentation of issues, proposed solutions, and any agreed changes to scope or priorities.
Track and report hours/effort weekly in accordance with time-reporting requirements.
Experience, Education, and Certifications:
Bachelor's degree in Computer Science, Data Science, Engineering, or a related field, or equivalent work experience.
Demonstrated experience designing or operating machine learning data pipelines, including dataset versioning, storage, and…