Senior AI Engineer

We need to output HTML fragment with sections: About the role, Responsibilities, Requirements, Qualifications, Skills, Benefits, Pay, Schedule. Only include sections that have real specific content. The posting includes description about role, responsibilities, requirements, qualifications, preferred qualifications. No explicit Benefits, Pay, Schedule. So we should include About the role (maybe from intro paragraphs), Responsibilities (list), Requirements (list), Qualifications (maybe combine required and preferred? Could separate but we have only one Qualifications section; we could include both required and preferred under Qualifications). Skills maybe part of Requirements. But we can follow instruction: Use headings for sections present. We'll include About the role, Responsibilities, Requirements, Qualifications. No Benefits/Pay/Schedule. We must not repeat job title as heading. Use

for each section heading. Use

• for any list of duties/requirements/benefits. Use

  for prose. We need to preserve all substantive information, not invent or summarize away detail. Remove boilerplate: EEO, etc. Also remove navigation, cookie, etc. Also remove duplicated lines. We need to extract content. Let's parse posting: Intro: "Austin, TX At IPT Associates (IPTA), we enjoy solving real-world problems with technology. We work closely with our customers, teammates, experts, and partners to come up with practical solutions that make a difference. Whether it's a government or business organization, we focus on understanding the need and building something that works. As a Senior AI Engineer, you are at the heart of our team's AI mission. You will architect, build, and deploy production-grade Generative AI applications, autonomous agent workflows, and retrieval-augmented generation (RAG) systems for Army decision support products. This is not a traditional ML modeling role, nor is it merely an LLMOps coding position. We require an engineer who possesses a deep theoretical and practical understanding of Foundation Models (FMs), reasoning architectures, and the mechanics underpinning modern AI systems. You will focus on operationalizing FMs, wiring deterministic software systems to non-deterministic LLM reasoning engines, and integrating these nodes directly into data transformation, visualization, and business process pipelines. Working alongside a dedicated mission team, you will take ownership of the AI architecture and AI Ops for your projects to ensure our systems meet strict operational standards where accuracy is the highest priority, followed by speed, and then cost. Our Team At IPTA, everything starts with our people. We're big believers that when you invest in your team, great things happen. That's why we encourage learning, growth, and trying new things-even if you don't have all the answers yet. Our culture is rooted in fierce determination, fearless integrity, and passionate service-but we also like to keep things collaborative, supportive, and down-to-earth. We're Looking For People Who Are curious and excited about technologyLike solving problems and figuring things outCan take initiative but also enjoy working with a teamWant to keep learning, growing, and challenging themselves If you're someone who's eager to jump in, learn something new, and make an impact-you'll fit right in here." We can combine into About the role paragraph(s). Probably keep as prose. Responsibilities: list bullet points under "Responsibilities" heading. Responsibilities bullet list: - Agentic System Architecture: Design and orchestrate multi-step autonomous AI agents primarily using Palantir AIP. Apply deep theoretical knowledge to implement reliable function calling, state machines, and human-in-the-loop (HITL) checkpoints - Theoretical & Applied Architecture: Serve as the team's intellectual lead on modern AI systems. Translate the underlying theory of attention mechanisms, vector embeddings, and reasoning frameworks into practical, production-ready software architecture - Deterministic-Stochastic Bridging: Architect resilient stateful backends that safely wrap probabilistic LLM outputs with deterministic validation, retry policies, strict JSON/Pydantic schemas, and structured error handling - RAG & Pipeline Integration: Build and scale production RAG pipelines across structured and unstructured Army datasets. Implement hybrid search, advanced chunking, metadata filtering, and re-ranking layers to feed directly into decision support products and visualizations within Foundry - AI Ops & Evaluation (LLMOps): Take complete ownership of project LLMOps. Implement automated regression testing and continuous evaluation pipelines for LLM outputs (e.g., faithfulness, semantic relevancy, hallucination detection) to guarantee deterministic accuracy - Safety, Security & Governance: Enforce guardrails around context boundaries, data isolation, multi-tenant RBAC, PII redaction, and mitigation against prompt injection/model evasion within a secure DoD environment - Performance & Cost Optimization: Manage context window budgets, token consumption metrics, latency profiling, and model routing strategies, strictly adhering to the priority matrix of Accuracy > Speed > Cost Requirements: list bullet points under "Requirements" heading. Requirements bullet list: - Security Clearance: Must hold a current, active DoD Secret clearance - Theoretical & Systems Depth: Deep theoretical and practical understanding of Generative AI architectures, foundation models, context management, and advanced prompt engineering (e.g., Chain-of-Thought reasoning) - Enterprise AI Experience: Proven experience building ontology-driven workflows or AI pipelines within Palantir Foundry, Palantir Maven, or other mature production systems. (Note: Extensive Army Vantage experience is not strictly required if you possess strong analogous production experience) - Programming & Systems Integration: 5+ years of experience with advanced proficiency in Python and TypeScript, featuring strong software engineering fundamentals (OOP, microservices, async programming, API design) - Agentic Framework Proof Points: Strong working knowledge of frameworks such as LangChain and LangGraph. While Palantir AIP is the primary platform, experience with these frameworks serves as a critical proof point of your architectural capabilities and will support future needs - Vector Search & Data Engineering: Proven experience indexing and querying vector databases/extensions (e.g., pgvector, OpenSearch, Pinecone, Qdrant) alongside relational and document datastores - Evaluation & Observability: Practical experience using LLM tracing and evaluation tools (e.g., Langfuse, Arize Phoenix, Ragas, TruLens) to diagnose multi-turn agent failures and latency bottlenecks Preferred Qualifications: list bullet points under maybe "Qualifications" heading (or separate Preferred Qualifications). Since we only have one Qualifications section, we can include both required and preferred? But we already have Requirements section for required. Qualifications could include preferred qualifications. We'll put Preferred Qualifications under Qualifications heading. Preferred Qualifications bullet list: - Active Top Secret (TS) Clearance - Experience operating within DoD Impact Level 5/6 (IL5/IL6) secure network environments - Industry Certifications: Recognized applied AI certifications (e.g., Google Cloud Professional Machine Learning Engineer/Generative AI Developer, Claude Architect, AWS Certified Generative AI Developer - Professional, Microsoft Azure AI Engineer Associate) - Cloud & Infrastructure: Demonstrated experience consuming model endpoints over REST/gRPC, deploying asynchronous worker queues, and integrating serverless compute/microservices We should not include Benefits, Pay, Schedule as they are not present. We need to output only HTML fragment, no markdown fences. Let's craft:

  About the role

  ... (combined prose)

  Responsibilities

  • ...
  ...

  Requirements

  • ...
  ...

  Qualifications

  • ...
  ...
  Make sure to preserve all details exactly as in posting. Need to keep punctuation, parentheses, etc. Let's copy the prose exactly as in posting but maybe combine paragraphs. Ensure we don't lose any content. Original prose: "Austin, TX At IPT Associates (IPTA), we enjoy solving real-world problems with technology. We work closely with our customers, teammates, experts, and partners to come up with practical solutions that make a difference. Whether it's a government or business organization, we focus on understanding the need and building something that works. As a Senior AI Engineer, you are at the heart of our team's AI mission. You will architect, build, and deploy production-grade Generative AI applications, autonomous agent workflows, and retrieval-augmented generation (RAG) systems for Army decision support products. This is not a traditional ML modeling role, nor is it merely an LLMOps coding position. We require an engineer who possesses a deep theoretical and practical understanding of Foundation Models (FMs), reasoning architectures, and the mechanics underpinning modern AI systems. You will focus on operationalizing FMs, wiring deterministic software systems to non-deterministic LLM reasoning engines, and integrating these nodes directly into data transformation, visualization, and business process pipelines. Working alongside a dedicated mission team, you will take ownership of the AI architecture and AI Ops for your projects to ensure our systems meet strict operational standards where accuracy is the highest priority, followed by speed, and then cost. Our Team At IPTA, everything starts with our people. We're big believers that when you invest in your team, great things happen. That's why we encourage learning, growth, and trying new things-even if you don't have all the answers yet. Our culture is rooted in fierce determination, fearless integrity, and passionate service-but we also like to keep things collaborative, supportive, and down-to-earth. We're Looking For People Who Are curious and excited about technologyLike solving problems and figuring things outCan take initiative but also enjoy working with a teamWant to keep learning, growing, and challenging themselves If you're someone who's eager to jump in, learn something new, and make an impact-you'll fit right in here." Note there are some missing spaces: "technologyLike", "outCan", "teamWant". We should preserve as is? Probably we should keep exactly but maybe we can add spaces for readability? The instruction: preserve ALL substantive information; do NOT invent or summarize away detail. Changing spacing could be considered altering detail? It's minor formatting but could be considered changing. Safer to keep exactly as given. However the text seems to have missing spaces due to PDF extraction. We could keep as is. We'll output that paragraph inside

  . Now responsibilities list: need to keep each bullet exactly as given. Let's copy each bullet: - Agentic System Architecture: Design and orchestrate multi-step autonomous AI agents primarily using Palantir AIP. Apply deep theoretical knowledge to implement reliable function calling, state machines, and human-in-the-loop (HITL) checkpoints - Theoretical & Applied Architecture: Serve as the team's intellectual lead on modern AI systems. Translate the underlying theory of attention mechanisms, vector embeddings, and reasoning frameworks into practical, production-ready software architecture - Deterministic-Stochastic Bridging: Architect resilient stateful backends that safely wrap probabilistic LLM outputs with deterministic validation, retry policies, strict JSON/Pydantic schemas, and structured error handling - RAG & Pipeline Integration: Build and scale production RAG pipelines across structured and unstructured Army datasets. Implement hybrid search, advanced chunking, metadata filtering, and re-ranking layers to feed directly into decision support products and visualizations within Foundry - AI Ops & Evaluation (LLMOps): Take complete ownership of project LLMOps. Implement automated regression testing and continuous evaluation pipelines for LLM outputs (e.g., faithfulness, semantic relevancy, hallucination detection) to guarantee deterministic accuracy - Safety, Security & Governance: Enforce guardrails around context boundaries, data isolation, multi-tenant RBAC, PII redaction, and mitigation against prompt injection/model evasion within a secure DoD environment - Performance & Cost Optimization: Manage context window budgets, token consumption metrics, latency profiling, and model routing strategies, strictly adhering to the priority matrix of Accuracy > Speed > Cost Now requirements list: - Security Clearance: Must hold a current, active DoD Secret clearance - Theoretical & Systems Depth: Deep theoretical and practical understanding of Generative AI architectures, foundation models, context management, and advanced prompt engineering (e.g., Chain-of-Thought reasoning) - Enterprise AI Experience: Proven experience building ontology-driven workflows or AI pipelines within Palantir Foundry