Senior Applied Scientist

We deliver technology to some of the best startups and companies in the world through diverse empowered teams of technologists that want to create change in the digital world.We are an organization that was created to be a safe place for all people from all areas of life. A place where teams are gender-balanced and compensated equally. And a place where careers are unrestrained regardless of cultural, educational, or geographical background.

We value people with strong technical skills that are collaborative, curious, results-driven, and take ownership. We embrace people that want to be themselves, have daily flexibility, grow, learn, and make a difference wherever the opportunity presents itself.

So we hope this sounds like you. Because we are always looking for exceptional Senior Applied Scientist Engineers to work in immersive client projects that will challenge your abilities. This position requires:

Role Summary

We are seeking a Senior Applied Scientist to design and deploy AI systems and agentic workflows that safely automate document understanding and decision-making across FOIA, Privacy Act, eDiscovery, and Data Breach workflows.
This role focuses on applied AI for high-risk, regulated documents, where systems must make or assist decisions related to review, classification, entity extraction, redaction, and disclosure, while remaining explainable, policy-aware, auditable, and defensible. You will work above core OCR and ingestion pipelines, transforming document content, metadata, layout, policy rules, and reviewer behavior into bounded automation that can be trusted in production. You will help define what decisions can be automated, under what conditions, and with what safeguards.

What You’ll Own

Automation of Document Decisions
Design AI systems that safely automate or assist decisions related to:
• Document review, including:
- Responsiveness and relevance
- Privilege and confidentiality
- Sensitivity and disclosure risk
• Entity and data extraction, including:
- Identifying, normalizing, and consolidating people, organizations, locations, and attributes
- Resolving entities across documents and datasets
• Redaction decisions, including:
- What information requires protection
- Where and how it should be applied
- Under which legal or policy authority

Define automation boundaries and confidence thresholds, including:
• Auto-apply vs. recommend vs. require human review
• Document-, page-, and entity-level decision logic
• Conditions under which automation must defer to reviewers
Establish human-in-the-loop controls so all automated decisions are reviewable, reversible, and auditable.

Applied GenAI, Agentic Workflows & Document Intelligence
• Build LLM-powered workflows for:
- Document and case summarization with citations and traceability
- Review assistance for responsiveness, privilege, and confidentiality
- Redaction and extraction rationale generation (the “why” behind decisions)
• Design and deploy Retrieval-Augmented Generation (RAG) systems grounded in:
-Document text
-Metadata and layout signals
-Entity models and prior reviewer decisions
• Design agentic AI workflows that orchestrate multi-step processes (e.g., ingest → classify → extract → review → redact → QA), with:
-Policy-aware reasoning
-Confidence gating
-Human approval loops
• Implement guardrails and grounding to ensure GenAI outputs remain accurate, explainable, and policy-aligned

Policy-Aware AI Systems
• Translate legal and regulatory requirements into clear, testable AI behavior
• Align systems with:
- FOIA exemptions and disclosure rules
-Privacy Act requirements
-eDiscovery standards for responsiveness, privilege, and confidentiality
- Data breach impact assessment and notification workflows
• Partner with legal, privacy, and compliance stakeholders to validate and refine decision logic

Quality Assurance, Trust & Defensibility
• Build QA frameworks for automated decisions, including:
-Review consistency and conflict detection
-Entity completeness and normalization checks
-Redaction over- and under-coverage analysis
• Monitor model performance, drift, and automation outcomes in production
• Ensure all decisions are traceable, explainable, and defensible for audits, litigation, and public scrutiny

Cross-Functional Leadership
• Collaborate with Product, Engineering, Legal, Privacy, and Operations teams
• Influence product direction across Legal, Breach, and Government markets
• Own AI outcomes in production—not just model metrics or prototypes

Required Qualifications

• Bachelor’s or Master’s degree in Computer Science, AI, Data Science, or a related field
• 5+ years of experience building and deploying applied ML / NLP systems in production, particularly for document understanding, classification, or extraction
• 2+ years of hands-on experience with large language models or modern GenAI techniques, including prompt design, RAG pipelines, or LLM-assisted workflows
• Strong Python proficiency and experience with frameworks such as:
o PyTorch or TensorFlow
o HuggingFace, spaCy, LangChain (or equivalent)
• Experience building explainable, auditable AI systems
• Experience designing human-in-the-loop automation
• Comfortable working in ambiguous, high-risk environments
• U.S. Citizen eligible for Public Trust clearance

Preferred Qualifications

• Experience with FOIA, Privacy Act, HIPAA, GDPR, or breach response workflows
• Background in eDiscovery, legal tech, or regulated government systems
• Experience with:
o Layout-aware document models
o OCR confidence and degradation analysis
o Large-scale entity resolution across document collections
• Prior technical leadership or mentoring experience
• Experience scaling AI systems to millions of documents or pages

A happy team makes a huge difference, that's why we provide:

• Payment in USD or in your local currency

• A truly flexible work schedule

• Holiday and performance bonuses

• An excellent paid time off policy

• 4 free Udemy courses a year

• Home exercise & wellness membership

• An opportunity for you to help create change in the industry

• And more!

ATHENAWORKS is an inclusive safe organization that only considers your technical ability, work experience, ability to collaborate, your capacity to grow to the next level of your career, and ability to deliver great work. This means that we also embrace/welcome self-taught people as well! We will NEVER consider any other personal or professional aspects of your life. We hope that you choose to have a conversation with us today and find out what makes us different from any company that you have experienced.