Data Scientist
Company: Obviant
Location: Arlington, VA (Remote)
Salary: $140,000 - $160,000 a year
Type: Full-time
Remote: Yes
Posted: 2026-11-17
About this role
About the Role:
As a Data Scientist at Obviant, you will turn thousands of messy, fragmented, and unstructured defense acquisition sources into clean, predictive, and scalable intelligence. You will lead high-impact data science and modeling initiatives designing statistical models, entity resolution pipelines, and automated analytical workflows that bring clarity to noisy datasets. Acting as a technical builder and strategic partner, you will replace manual, ad-hoc analysis with resilient, production-grade models and automated decision systems built to scale alongside our growing data footprint.
What You’ll Do
- **Tame Messy, Unstructured Data:** Design and deploy automated pipelines for data cleaning, entity resolution, deduplication, and feature engineering across disparate, noisy public-sector and commercial datasets.
- **Statistical & Predictive Data Modeling:** Architect, validate, and productionize scalable data models—spanning forecasting, classification, anomaly detection, and network/graph modeling—to surface actionable defense market intelligence.
- **End-to-End Automation:** Audit manual analytical workflows and ad-hoc exploratory scripts, refactoring them into automated, reproducible, and self-healing production pipelines that run at scale.
- **Build for Scale & Reliability:** Partner closely with Data and Backend Engineering to ensure data models, feature stores, and automated scoring jobs handle rapidly expanding data volumes with low latency and high fault tolerance.
- **Automated Data Quality & Observability:** Pioneer automated validation frameworks, drift detection, and statistical monitoring to ensure model accuracy and data integrity as new sources are ingested.
- **Cross-Functional Impact & Translation:** Collaborate with Product, Engineering, and leadership to frame ambiguous business and government workflows as tractable data science problems, translating complex model outputs into clear, trustworthy product features and strategic ...