Data Architect Engineer

Tria Federal•Washington, DC

3d•Onsite

About The Position

A Senior Data Architect Engineer to support the DIA O&I Enterprise Integration and Assessments Data Team, focused on accelerating dataset understanding and delivering fit-for-purpose machine learning and data solutions. The role will help the team move from “what data do we have and what can it support?” to “what model or workflow should we implement, and how do we sustain it?” The emphasis is on selecting the right method or tool for the mission problem, including determining when ML is appropriate and when rules, heuristics, or simpler analytics are better suited.

Requirements

B.S. in related field with 5+ years of experience or 10+ years with no B.S.
Must be a U.S. Citizen
Must have an ACTIVE TS/SCI w/ CI Poly Clearance
Senior-level experience supporting machine learning, data science, data architecture, or data engineering efforts
Strong Python and SQL experience
Experience with common machine learning frameworks and libraries, such as scikit-learn, PyTorch, or TensorFlow
Experience assessing datasets for ML readiness, including data quality, metadata, labeling, ground truth, feature engineering, and constraints
Experience designing, building, evaluating, and improving ML models based on mission or business use cases
Experience defining model evaluation metrics and conducting error analysis
Experience building reproducible technical workflows and documenting implementation approaches
Experience using AWS SageMaker for experimentation, training, processing, pipelines, model registry, or deployment approaches
Sound software engineering practices, including Git, readable and modular code, basic testing, documentation, and reproducibility
Ability to work independently, provide technical recommendations, and interface with Government leads and senior stakeholders with limited oversight

Nice To Haves

Familiarity with scalable data processing tools such as Spark or PySpark is a plus

Responsibilities

Independently assess and explain datasets, including content, structure, quality, gaps, lineage, metadata, constraints, and known limitations
Identify what is needed to make datasets ML-ready, including labeling, ground truth, feature creation, and data quality considerations
Recommend practical approaches for addressing risk considerations such as bias, drift, and model limitations
Design, build, and iterate ML models appropriate to the data and use case, such as classification, entity or record matching, anomaly detection, and natural language processing, as applicable
Establish baselines and define evaluation metrics tied to operational utility
Perform error analysis to guide model improvements and inform recommendations
Build reproducible workflows that can be rerun and sustained within the customer’s operating environment and security constraints
Support implementation decisions, including batch versus real-time processing, resource and cost tradeoffs, and latency or throughput considerations
Use AWS SageMaker for experimentation and execution, including notebooks or Studio, training jobs, processing jobs, pipelines or automation, model registry, and deployment approaches as permitted
Provide clear technical recommendations to the team and Government lead on data strategy, modeling choices, architecture patterns, and implementation plans
Proactively identify technical risks, opportunities, and recommended paths forward with minimal oversight