Principal ML Engineer
signify
Job Description
- Build and operate GenAI and agentic AI systems using RAG, embeddings, vector search, prompt chaining, tool-using agents, multi-agent orchestration and guardrails.
- Lead ML engineering and MLOps for Advanced Analytics, including architecture, delivery standards, reusable patterns, platform adoption and production support.
- Own productionization of ML and GenAI solutions across forecasting, marketing mix modeling, customer segmentation, contract intelligence, IoT analytics and enterprise copilots.
- Design and implement scalable ML pipelines across data ingestion, feature engineering, training, evaluation, deployment, monitoring, retraining and model retirement.
- Define target architecture for enterprise GenAI platforms, including agent registry, MCP registry, agent-to-agent communication, observability, safety, lifecycle management and reusable integration patterns.
- Engineer reliable agent runtimes with retry logic, timeout handling, fallback strategies, deterministic execution where required, idempotency, execution tracing and production SLOs.
- Design and own evaluation frameworks for ML, GenAI and agentic AI systems, including offline and online evals, task success metrics, hallucination checks, latency, cost, quality and regression benchmarks.
- Establish LLMOps practices covering prompt testing, model evaluation, data curation, fine-tuning workflows, safety testing and continuous improvement from production feedback.
- Optimize GenAI workloads for latency, throughput and cost using model routing, caching, batching, autoscaling, token usage monitoring and fit-for-purpose model selection.
- Build reusable AI platform services, APIs, microservices and integration patterns that allow internal teams to safely consume ML and GenAI capabilities.
- Create observability dashboards for model performance, data drift, agent traces, token usage, latency, reasoning paths, production failures and business impact.
- Integrate ML and GenAI workloads with Snowflake using Snowpark, APIs, storage integrations and Cortex where relevant.
- Use AWS services such as SageMaker, Bedrock, Lambda, Fargate, EC2, S3, SNS and SQS to deliver reliable cloud-native ML platforms.
- Establish CI/CD practices for ML, model versioning, automated testing, reproducible deployments, data quality checks, access control and deployment readiness gates.
- Own experimentation infrastructure, including A/B testing, offline evaluation and production feedback loops for ML and GenAI products.
- Guide ML engineers and data scientists on architecture, coding practices, experimentation, model evaluation, deployment readiness and operational excellence.
- Partner with business, digital, CDIO and analytics stakeholders to translate business problems into scalable AI products and measurable outcomes.
- Evaluate emerging platforms and tools such as Snowflake Cortex, Bedrock Agents, open-source LLMs, agent frameworks, MCP-based integrations and AI observability platforms, and recommend adoption paths.
- Champion responsible AI practices covering explainability, fairness, transparency, auditability, access control, privacy, security and safe use of enterprise data.
- Present technical designs, trade-offs, risks and business outcomes clearly to senior stakeholders.