Data Scientist

Peraton•Red Bank, NJ

About The Position

We are seeking a Data Scientist to build and curate the knowledge base for a generative AI/ML research effort, and to develop the data and knowledge extraction pipelines for machine learning models. The role is responsible for extracting, structuring, and validating design knowledge into knowledge graphs and ontologies; for engineering the metadata, annotation, and provenance of program datasets; and for the analysis that turns results, laboratory measurements, and simulation output into actionable findings for the research team. The ideal candidate is a strong applied data scientist with an interest in symbolic and structured representations of knowledge. Experience with formal methods and domain-specific languages (DSLs) is desired but not required.

Requirements

Bachelor’s Degree or higher in Computer Science, Statistics, Mathematics, Electrical Engineering, or a related technical field
5+ years of applied data science or data engineering experience (or MS with 3+ years) with a record of delivering data pipelines and analyses that other engineers and researchers depend on
Strong Python proficiency including the scientific stack (NumPy, pandas, SciPy, scikit-learn) and experience with at least one deep learning framework (PyTorch preferred)
Hands-on experience with knowledge graphs, ontologies, or structured knowledge representation (RDF/OWL, property graphs such as Neo4j, or equivalent) and with querying and validating them (SPARQL, Cypher, SHACL, or similar)
Experience designing data schemas, metadata standards, and annotation workflows for large scientific or engineering datasets, including data versioning and provenance
Solid statistical foundations: experimental design, hypothesis testing, uncertainty quantification, and the ability to explain results to technical and non-technical audiences
Experience with SQL and with at least one workflow or pipeline orchestration tool (Airflow, Prefect, Dagster, DVC, or equivalent)
Ability to produce clear documentation including data dictionaries, dataset cards, and analysis reports
US Citizenship

Nice To Haves

Experience with formal methods or formal verification, such as SMT solvers (Z3, cvc5), model checkers, property-based testing (Hypothesis), or proof assistants, particularly applied to validating generated programs or signal processing pipelines
Experience designing or implementing domain-specific languages: grammar design, parser generators (ANTLR, Lark, or similar), type systems, intermediate representations, or compiling DSL programs to executable code
Background in symbolic AI or neuro-symbolic methods: logic programming, constraint solving, rule engines, or program synthesis
Familiarity with digital signal processing and communications fundamentals (modulation, filtering, coding, channel effects) or with GNU Radio and software-defined radio data formats (I/Q sample handling, SigMF or similar metadata standards)
Prior work on IARPA, DARPA, or similar government research programs, including data sharing plans, privacy protection plans, and delivery of datasets to independent T&E teams
Experience building causal or probabilistic models from structured knowledge, or working alongside causal inference researchers
MS or PhD in Computer Science, Statistics, Electrical Engineering, or a related technical field
Willingness and ability to obtain Secret security clearance

Responsibilities

Design, build, and populate the research team’s effort knowledge bases, and maintain it under version control with provenance tracking
Develop knowledge extraction pipelines (rule-based, statistical, and ML-assisted) that convert unstructured and semi-structured sources into structured, queryable knowledge; establish quality metrics and validation procedures for extracted content
Engineer the metadata, annotation schema, and packaging for program datasets (synthetic, simulated, and real collections) so that datasets are reproducible, well documented, and deliverable on the program data sharing schedule
Perform exploratory and inferential analysis on training data, simulation output, and evaluation results; build dashboards and reports that show where generated waveforms succeed or fail against objectives, and feed findings back to AI model researchers and engineers
Contribute data and analysis sections to design reviews, monthly status reports, and dataset documentation; coordinate with academic subcontractors on shared data and knowledge resources