Data Scientist / ML Engineer

Cognizant is hiring for an onsite role in Phoenix, AZ focused on building production ML/AI capabilities that improve how entities are resolved into golden records and how ontology-aligned intelligence is applied. You will deploy and iterate models on Vertex AI, using Gemini APIs for labeling support and working closely with the team to validate results against ontology rules.

Compensation: USD 110,000 - 125,000 per year. Experience: 5+ years.

What you’ll get

• Medical/Dental/Vision/Life Insurance
• Paid holidays plus Paid Time Off
• 401(k) plan and contributions
• Long-term/Short-term Disability
• Paid Parental Leave
• Employee Stock Purchase Plan

Responsibilities

• Own entity matching and bridging by designing and building fuzzy and exact matching models to stitch duplicate entities into golden records.
• Build inference models to support data profiling, including identifying patterns, relationships, and anomalies in profiled source data.
• Advance the business-ontology effort using Gemini-assisted labeling models that classify records into business domains (Location, Asset, Service, Event, Work-ticket).
• Design and implement labeling workflows that combine weak supervision, LLM-assisted labeling, and human-in-the-loop review.
• Deploy and manage matching, inference, and labeling models on Vertex AI.
• Evaluate performance using precision/recall and drift, and collaborate with the Architect and Data Domain SME to validate outputs against ontology rules.
• Partner with the Backend/Agent Engineer to expose matching and labeling models as callable tools within the agent orchestration layer.
• Continuously retrain and tune models as new source data and golden-record feedback become available.
• Document model logic, training data lineage, and known limitations for each deployed model.
• Support QA by defining acceptance thresholds for model-driven outputs such as matching accuracy and labeling precision.

Requirements

• 5+ years in applied ML, including entity resolution, classification, or NLP-based matching.
• Strong Python experience with PyTorch and/or scikit-learn.
• Hands-on experience deploying and evaluating models on Vertex AI.
• Experience using embeddings, LLMs, or NLP techniques for matching and classification.
• Experience designing labeling workflows using weak supervision, LLM-assisted labeling, and human-in-the-loop review.

Preferred qualifications

• Experience with knowledge graphs or ontology-aligned ML (feature use over graph data).
• MLOps experience including model versioning, monitoring, and retraining pipelines.
• GCP ML Engineer certification.

Core skills

• Entity resolution and matching: fuzzy and exact matching for golden-record creation.
• Inference and pattern-detection over profiled data.
• LLM-assisted labeling and classification (Gemini) with human-in-the-loop workflow design.
• Model deployment, evaluation, and drift monitoring on Vertex AI.
• Weak supervision and semi-supervised labeling workflow design.

Technologies

• Vertex AI
• Gemini APIs
• Python
• PyTorch
• scikit-learn
• BigQuery
• Spanner Graph

Role summary

Build the matching, inference, and labeling models that resolve entities into golden records and bring Gemini-assisted intelligence into the ontology.