Artificial Intelligence and Machine Learning Engineer

RESULTS. INNOVATION. VALUES. ACCOUNTABILITY.

That's RIVA.

We're a mission-driven IT services company and systems integrator supporting digital transformation and modernization for federal government agencies. Since 2009, we've partnered with our customers to solve complex challenges through smart, practical innovation to deliver real outcomes where they matter most. Our teams are made up of industry-leading experts who are passionate about doing great work and making a difference. We don't just develop solutions, we support efforts that strengthen communities and serve the public good.

RIVA's culture is built on four core values: Results, Innovation, Values, and Accountability. They guide how we work, how we collaborate, and how we measure success. Our employee-first approach is rooted in trust, ownership, and meaningful work. By investing in our people and fostering a flexible, supportive environment, employees have the opportunity to grow their skills, contribute ideas, and make an impact from day one. all while supporting missions that matter.

POSITION OVERVIEW

RIVA Solutions seeks an AI Subject Matter Expert to support the USPTO P-TACTS program, the Patent Trial and Appeal Board (PTAB) case tracking and electronic filing system. You will design, build, and evaluate production AI solutions that turn unstructured

legal filings into reliable, structured data. A current example: a pipeline that reads PTAB petitions, institution decisions, and final written decisions, then extracts trial-level and claim-level outcomes. The team measures the pipeline against analyst-labeled ground truth. The output supports reporting that PTAB delivers to Congress. You will work mainly in Python on Databricks, inside a secured federal environment. Development runs on remote instances through VPN. LLM access runs through an agency AI gateway, and approved models trail the commercial market. You will work directly with government stakeholders and engineers to turn mission needs into working, maintainable, and measurable applications.

WHAT YOU WILL DO

• Build LLM pipelines that extract structured data from long, unstructured legal documents.
• Decide when a task needs a flexible agentic harness and when it needs a constrained, step-by-step pipeline. Build either one.
• Validate model outputs against schemas, allowed values, and business rules. Handle failures without losing records silently.
• Design evaluations against ground truth. Choose the right metric for the problem, run error analysis, and report results in plain terms.
• Use AI coding tools such as Claude Code every day to build, test, and review code.
• Keep pipelines stable as approved models change, using version pinning, regression test sets, and prompt versioning.
• Advise the customer on future use cases, including fine-tuning small language models or adapters when the data supports it.

MINIMUM QUALIFICATIONS

• Bachelor's degree in Computer Science, Data Science, Engineering, or a related field.
• 5+ years of applied ML, data science, or AI engineering experience, including 2+ years building applications on large language models in production or near-production settings.
• Strong Python skills, including data work with pandas or PySpark.
• Solid ML fundamentals: classification, train and test splits, overfitting, data leakage, and model selection.
• Hands-on evaluation skill: precision, recall, F1, confusion matrices, class imbalance, and comparison to labeled ground truth.
• Production experience building with LLM APIs, including structured output, tool calling, and prompt design.
• Experience building or debugging an agent harness, and the judgment to explain when a structured pipeline is the better choice.
• Experience validating LLM output with schemas (for example, Pydantic or JSON Schema), retries, and exception handling.
• Daily use of AI coding assistants such as Claude Code, Codex, Cursor, or GitHub Copilot.
• Hands-on experience building and running ML or LLM workloads on Databricks or a major cloud platform (AWS, Azure, or GCP). Examples: data pipelines in PySpark or notebooks, calling hosted models through Databricks Model Serving, Amazon Bedrock, Amazon SageMaker, Azure OpenAI, Azure Machine Learning, or Vertex AI, and tracking experiments in MLflow.
• AWS Certified Machine Learning Engineer - Associate, or an equivalent cloud ML certification. An active AWS Certified Machine Learning - Specialty also counts.
• Eligibility to obtain a USPTO Public Trust and authorization to work in the U.S.

PREFERRED QUALIFICATIONS

• Databricks experience, including Jobs, Delta Lake, MLflow, and Unity Catalog. PyTorch experience, including fine-tuning small language models or training LoRA adapters.
• AWS experience, including Amazon Bedrock.
• Information extraction from scanned or long documents, including OCR and layout analysis.
• Retrieval augmented generation and vector databases.
• MLOps

experience

pipeline orchestration, evaluation sets run in CI, and monitoring of output quality over time.

• Experience working through an enterprise AI gateway with cost tracking and request filtering.
• Databricks Certified Machine Learning Professional, Databricks Certified Machine Learning Associate, Databricks Certified Generative AI Engineer Associate, AWS Certified Machine Learning Engineer - Associate, or an equivalent cloud certification.
• Experience with USPTO, PTAB, or patent and legal data.
• Prior delivery on federal government contracts, including FISMA Moderate systems.

EQUAL EMPLOYMENT OPPORTUNITY & ACCOMMODATION

We believe great teams are built from different backgrounds, perspectives, and lived experiences, and we mean that beyond the buzzwords.

RIVA is an equal opportunity employer. We welcome applicants of every race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, veteran status, and any other characteristic protected by law.

If you need an acco