Data Scientist, Private AI/Big Data

Data Scientist, Private AI/Big Data at Apoddo

We’re looking for a Data Scientist to work side-by-side with a Senior AI Engineer building AI agents inside a private-infrastructure SaaS platform that works with 100,000 records on a daily basis.

What You’ll Do

• Define Golden Record logic: Specify how to resolve conflicts within a duplicate cluster (which record is correct, which is stale/erroneous, and what should be merged).
• Design entity resolution strategies: Build and evaluate matching approaches (rules + ML), including blocking/candidate generation, pair scoring, and cluster-level decisions.
• Build labeling + evaluation pipelines: Create ground-truth datasets and annotation guidelines; measure precision/recall, cluster purity, and “bad merge” risk.
• Develop features & signals: Engineer features from structured data (names, addresses, timestamps, metadata, relationships across tables) and from embeddings/vector search signals.
• Reduce risk with confidence + explainability: Provide confidence scoring, evidence summaries, and decision reasoning to support human-in-the-loop verification.
• Partner with AI Engineering: Collaborate on agent prompts/workflows, retrieval strategies, and safe execution checkpoints before UPDATE/DELETE operations.
• Monitor production quality: Define metrics, drift detection, and feedback loops so the system improves over time.

Key Tech & Data Environment

• Datastores: PostgreSQL, Elasticsearch (vector/kNN search)
• AI Workflows: LangGraph (stateful reasoning), agent orchestration patterns
• Models (Private): Llama 3.x / Mistral via vLLM or Ollama
• Infra: Kubernetes, Docker, private AWS/Azure environments

Who You Are

• 3+ years (or equivalent) experience as a Data Scientist / ML Engineer / Applied Scientist working with real-world messy data.
• Strong grasp of entity resolution / record linkage / deduplication concepts (matching, blocking, clustering, merge strategies).
• Comfortable with SQL and large-scale data analysis; can navigate multi-table relational schemas confidently.
• Experience designing evaluation frameworks and datasets (labeling strategy, sampling, measuring trade-offs).
• Pragmatic mindset: you care about false merges as much as false negatives, and you know how to reduce operational risk.
• Strong communication skills—able to turn ambiguous data issues into clear rules, metrics, and experiments.

Nice to Have

• Experience with customer/account/event data domains (CRM, transactions, identity, behavioral events).
• Familiarity with vector search, embeddings, or retrieval-based matching.
• Experience collaborating on LLM/agent systems (RAG, tool use, human-in-the-loop validation).
• Background in data quality, MDM, or data governance.

Key skills/competency

• Entity Resolution
• Machine Learning
• Data Quality
• Feature Engineering
• Evaluation Frameworks
• LLM/AI Agents
• Big Data Analysis
• SQL
• Production Monitoring
• Cloud Infrastructure