AI Engineer

• ? Build production applications using large language models, smaller task-specific models, retrieval-augmented generation, tool/function calling, workflow orchestration, and deterministic business logic where appropriate.
• ? Develop secure APIs, services, adapters, and event-driven integrations for digital channels, customer-care platforms, enterprise knowledge, billing and entitlement services, content/metadata systems, and internal workflows.
• ? Implement authorization-aware tool use, input validation, idempotency, timeouts, retries, fallback behaviour, circuit breakers, and human escalation paths.
• ? Choose prompts, retrieval, rules, conventional machine learning, or fine-tuning based on evidence rather than defaulting every problem to a large model.

Retrieval, data, and grounding:

• ? Build ingestion, chunking, metadata, indexing, retrieval, reranking, citation, freshness, and deletion workflows for enterprise knowledge and approved content sources.
• ? Preserve source permissions and customer/data boundaries throughout retrieval and generation; prevent unauthorized cross-user, cross-account, or cross-domain disclosure.
• ? Partner with Data Engineering and domain owners on data quality, system-of-record alignment, lineage, and feedback loops. Evaluation and quality engineering
• ? Create representative evaluation datasets and automated test suites for groundedness, relevance, correctness, task completion, refusal behaviour, safety, robustness, latency, and cost.
• ? Run regression testing across prompt, model, retrieval, tool, and policy changes; analyse failure modes and improve the system using trace-based evidence.
• ? Instrument online quality and business metrics, support controlled experiments, and incorporate human review for higher-risk or lower-confidence outcomes. Production operations and MLOps
• ? Build CI/CD pipelines for code, configuration, prompts, evaluation assets, and model or index changes across separated development, test, and production environments.
• ? Implement structured logging, tracing, token and infrastructure cost monitoring, model/provider health checks, alerting, dashboards, and operational runbooks.
• ? Optimize throughput, latency, reliability, and cost using caching, batching, routing, prompt/context management, and appropriately sized models.
• ? Participate in production support, incident response, root-cause analysis, and continuous improvement.

Security and responsible implementation:

• ? Implement controls for prompt injection, jailbreak attempts, unsafe tool use, data leakage, malicious content, model abuse, and dependency/supply-chain risk.
• ? Apply DIRECTV requirements for PII and payment-card data, identity and access, secrets management, retention, content rights, audit logging, and approved model/provider use.
• ? Contribute reusable components to the AI control plane, including policy enforcement, prompt/model configuration, evaluation hooks, telemetry, and kill-switch or rollback mechanisms.
• ? Work with Product Managers, UX, Solution Architects, AI Architects, Data Engineers, Cybersecurity, Quality Engineering, and Operations to deliver testable user outcomes.
• ? Write maintainable code, automated tests, interface contracts, technical documentation, deployment guides, and operational runbooks; participate in code and design reviews.

• Typically, 10+ years in professional software engineering, including meaningful hands-on experience delivering AI, machine-learning, search, NLP, or data-intensive applications to production; equivalent experience is welcome.
• Hands-on experience with LLM APIs, prompt and context design, RAG, embedding/search systems, structured outputs, tool/function calling, and automated evaluation.
• Experience with SQL and document/search stores, containers, CI/CD, source control, cloud services, and observability practices.
• Strong software engineering habits: modular design, automated testing, secure coding, peer review, performance troubleshooting, and production ownership.
• Ability to explain model limitations and engineering trade-offs to technical and nontechnical partners.
• Bachelor’s degree in computer science, engineering, data science, or a related field, or equivalent practical experience.