DE&A - AIML - Auto ML

Senior GenAI Data Engineering Developer (AI/GenAI Architect – Hands-on)

8–12 years (Data/AI Engineering), with 2+ years in AI/GenAI architecture and solution design

We’re seeking a hands-on GenAI Data Engineering leader who can architect and build production-grade GenAI solutions—from data pipelines and vectorization to RAG, LLM orchestration, governance, and cost-aware operations. You will translate business problems into secure, scalable, and compliant AI systems using LLMs, embeddings, and modern data stacks across Azure/AWS/GCP.

• Lead end-to-end GenAI solution architecture (Assess → Design → Build → Operate) for use cases like RAG/Q&A, copilots, summarization, classification, agents, autonomous workflows.
• Define LLMOps/MLOps blueprints: environments, CI/CD, model registry, observability, evaluation, A/B testing, canary rollouts, guardrails.
• Design RAG architectures: document loaders, chunking strategies, embeddings selection, vector schema design, re-ranking, caching, and fallbacks.
• Establish data governance for AI: PII handling, safety, red-teaming, content filters, model risk, usage policies, and auditability.

• Build robust ingestion & transformation pipelines (batch/streaming) to prepare high-quality corpora for LLMs.
• Operationalize chunking/embedding/vector indexing, metadata enrichment, synonyms/ontologies, and semantic retrieval performance tuning.
• Implement feature/knowledge stores, vector DBs, and document stores (e.g., Azure AI Search, Elasticsearch, Pinecone, Weaviate, Milvus, pgvector).
• Integrate orchestration frameworks (Airflow/Prefect/AKS/Databricks Jobs) and API gateways.

• Develop prompt pipelines (system/hybrid prompts, tool-use), retrieval chains, agents, function-calling, and tool integrations (SQL, search, APIs).
• Build and harden LLM applications (e.g., FastAPI/Flask/Functions) with authentication/authorization, rate-limiting, telemetry, and cost controls.
• Introduce guardrails (PII scrubbing, jailbreak mitigation, toxicity, hallucination checks) and evaluation harnesses (BLEU/ROUGE/METEOR, custom rubric scoring, human-in-the-loop).

• Set up model and app observability (latency, token usage, failure modes, retrieval quality, drift detection).
• Implement cost monitoring (per-call, per-user, per-use-case), prompt/embedding caching, and routing to optimize spend/performance.
• Drive SLA/SLO definitions, incident runbooks, and reliability engineering practices.

• Partner with Product, Security, Compliance, and Enterprise Architecture to align business outcomes with responsible AI.
• Lead technical design reviews, mentor developers, and contribute to standards/patterns across teams.

• GenAI Architecture & Delivery
  • Proven design/delivery of RAG and LLM apps in production
  • Expertise in prompt engineering, prompt templating, evaluation, guardrails
  • Experience with model selection (proprietary vs open-source), routing, and fallback strategies
• Data Engineering Excellence
  • Strong in Python (data processing, APIs, ETL/ELT, testing)
  • Proficient with SQL (analytical queries, performance tuning, stored procedures as needed)
  • Experience building scalable pipelines (Databricks/Spark, Airflow/Prefect, Kafka/EventHub)
• Vector & Retrieval Systems
  • Hands-on with vector databases and embedding pipelines
  • Mastery of chunking, retrieval optimization, re-ranking, metadata strategies
• Cloud & Platform
  • One or more: Azure (OpenAI, AI Search, Databricks, ADF/ADF v2/Synapse, AKS/Functions), AWS (Bedrock, OpenSearch, Sagemaker, Lambda/EKS), GCP (Vertex AI, BigQuery, GKE)
  • Containerization & CI/CD: Docker, Kubernetes, GitHub Actions/Azure DevOps/Jenkins
• Security, Governance & Compliance
  • Experience implementing Responsible AI, data privacy/PII, RBAC/ABAC, secret management, network isolation, policy-as-code
• Communication & Leadership
  • Ability to translate business problems to GenAI architectures and guide teams through delivery

• LLM Frameworks & Tools: LangChain, LlamaIndex, Semantic Kernel, DSPy
• Observability/Eval: MLflow, Promptfoo, TruLens, Arize, EvidentlyAI, OpenTelemetry, Kibana/Grafana
• Search/Retrieval: Elasticsearch/OpenSearch, Redis Stack, Vespa
• NLP/ML: Transformers, fine-tuning/LoRA, vector quantization, distillation
• Data Quality: Great Expectations/Deequ, Monte Carlo
• Edge/Hybrid: On-prem GPU, NVIDIA NIM, Triton Inference Server
• Compliance: SOC2, HIPAA, GDPR familiarity in AI contexts

• Bachelor’s/Master’s in Computer Science, Data Engineering, AI/ML, or related field
• 8–12 years in data/AI engineering; 2+ years in GenAI/LLM architecture
• Track record delivering secure, reliable, cost-efficient GenAI solutions at enterprise scale