Senior Data Scientist

Skywarditsolutions

Maryland

On-site

USD 110,000 - 145,000

Full time

14 days+

Get more replies from employers

Send a job-specific resume in minutes.

Benefits offered by this job

Medical, dental, and vision insurance

15 days of paid leave

401(k) with employer contribution

$5,000 for professional development

Job summary

Skywarditsolutions is seeking a Senior Data Scientist in Maryland to lead complex data science experiments and work with federal and private datasets. The ideal candidate must have extensive experience in graph analytics and strong skills in Python tools.

This role involves creating analyses, implementing advanced data techniques, and engaging with stakeholders to present findings. Offers include comprehensive benefits and a focus on exemplary data management and reproducibility.

Qualifications

7+ years of applied data science experience, with focus on graph analytics.
Real-world experience with probabilistic record linkage.
Strong production-grade Python skills.

Responsibilities

Lead end-to-end data science experiments and analyses.
Apply graph analytics and train interpretable classifiers.
Implement LLM-based named entity recognition.

Skills

Expertise in graph analytics

Strong Python skills

Experience with entity resolution tools

Anomaly detection techniques

Statistical analysis

Effective communication

Education

7+ years of data science experience

Active DHS Public Trust clearance

Tools

Python libraries (pandas, NumPy, scikit-learn)

Graph neural network libraries

Entity resolution tools (Senzing)

Job description

Job Overview

We need a Senior Data Scientist for a contingent hire based on contract award. The role requires expertise in navigating complex federal and private datasets that do not share schemas or IDs, and a focus on understanding data before modeling.

Responsibilities

Lead end-to-end data science experiments: assess data readiness, perform clustering and topological risk modeling, enrich with unstructured data, and resolve entity discrepancies.
Run exploratory data analyses (EDAs) on government-furnished data within a government‑controlled environment: profile completeness, identify schema mismatches, flag gaps, and document data limitations before any model is trained.
Apply graph analytics such as Leiden community detection, betweenness and eigenvector centrality, motif analysis, temporal cluster detection, and link prediction; explain results in plain English.
Train interpretable classifiers (logistic regression, gradient‑boosted trees) and Graph Neural Networks (GraphSAGE, GAT) where data allow; use unsupervised anomaly detection when labels are scarce.
Execute probabilistic entity resolution across biographic, behavioral, and biometric features using tools such as Senzing; handle name transliteration, DOB variation, and fuzzy address matching.
Implement LLM‑based named entity recognition and relationship extraction on unstructured field text; quantify whether extracted edges meaningfully alter the graph rather than add noise.
Wrangle messy data and recommend supplemental, de‑identified data sources that would enrich the analysis; document the case for each recommendation so privacy and legal teams can evaluate.
Document everything so the customer can rerun it later; reproducibility is a hard acceptance criterion, not a “nice to have.”
Maintain rigorous standards: provide confidence intervals on all outputs, record null findings with equal care, and avoid dashboard theater.

Qualifications

An active DHS Public Trust clearance at time of hire.
7+ years of applied data science experience, with at least 2 years focused on graph analytics or network analysis.
Strong production‑grade Python skills: pandas, NumPy, scikit‑learn, networkx (or graph‑tool/igraph), and at least one GNN library such as PyTorch Geometric or DGL.
Real‑world experience with probabilistic record linkage or entity resolution tools such as Senzing, dedupe, FEBRL, Magellan, or a defensible home‑grown Fellegi‑Sunter implementation.
Comfort working without labels: anomaly detection, positive‑unlabeled learning, isolation forests, autoencoders; know how to evaluate a model lacking clear ground truth.
Interpretability discipline: SHAP, feature importance, partial dependence plots, and the judgment to select simpler models when appropriate.
LLM application experience beyond prompt‑and‑pray: entity and relationship extraction on real text with genuine evaluation of extraction quality.
Statistical maturity: understanding of power analysis, confidence intervals, and avoidance of p‑hacking even when unchecked.
Ability to present technical findings to non‑technical operational stakeholders without losing nuance.
Reproducibility hygiene: version control, parameterized pipelines, deterministic seeds, pinned dependencies, and functional READMEs.
Published research, open‑source contributions, or conference talks in entity resolution, graph ML, or applied causal inference.
Knowledge of commercial data sources such as LexisNexis, TransUnion, Babel Street, etc.
A track record of explaining null results to federal subject matter experts without flinching.
Reputation among teammates for catching pipeline bugs before delivery.

Even if you don’t meet 100 % of the qualifications, we encourage you to apply. At Skyward, we’re focused on hiring individuals with the right skills and passion to grow, not just checking off every box.

Benefits