Design and maintain robust data ingestion and feature-store pipelines supporting both traditional ML models and large-language-model (LLM) applications. Implement Retrieval-Augmented Generation systems: build vector stores, design embedding strategies, integrate semantic search over document corpora, and manage latency/throughput trade-offs for real-time generation. Evaluate and optimize retrieval components for scale and relevance, and seamlessly integrate them with LLMs. Model and construct enterprise Knowledge Graphs: define ontologies, entity schemas, relation types, and ingestion pipelines (e.g. Neo4j). Develop graph-based inference and reasoning capabilities to enrich upstream data and to ground generative outputs in verified facts. Collaborate with data architects to ensure KG alignment with existing data warehouses, metadata catalogs, and governance policies. Architect goal-driven AI agents capable of multi-step planning and execution: break high-level objectives into tasks, sequence actions, handle failures, and loop back with new data. Integrate and extend orchestration frameworks such as Microsoft AutoGen, LangChain Agents, or custom task schedulers to manage agent workflows, tool-use, and state persistence. Implement secure, auditable agent pipelines with human-in-the-loop checkpoints, logging, and governance controls. Containerize and deploy ML and agentic components with Kubernetes or serverless platforms; build CI/CD for model, RAG-component, KG, and agent updates. Monitor system performance end-to-end: model drift, retrieval quality, graph consistency, agent accuracy, and overall SLA/SLI adherence. Automate alerting, rollback strategies, and canary-based rollouts to minimize runtime risk. Stay current on advances in RAG architectures, graph-based neural models, autonomous agent research, and orchestration tooling; propose POCs or pilot projects.
experience
6 ... Design and maintain robust data ingestion and feature-store pipelines supporting both traditional ML models and large-language-model (LLM) applications. Implement Retrieval-Augmented Generation systems: build vector stores, design embedding strategies, integrate semantic search over document corpora, and manage latency/throughput trade-offs for real-time generation. Evaluate and optimize retrieval components for scale and relevance, and seamlessly integrate them with LLMs. Model and construct enterprise Knowledge Graphs: define ontologies, entity schemas, relation types, and ingestion pipelines (e.g. Neo4j). Develop graph-based inference and reasoning capabilities to enrich upstream data and to ground generative outputs in verified facts. Collaborate with data architects to ensure KG alignment with existing data warehouses, metadata catalogs, and governance policies. Architect goal-driven AI agents capable of multi-step planning and execution: break high-level objectives into tasks, sequence actions, handle failures, and loop back with new data. Integrate and extend orchestration frameworks such as Microsoft AutoGen, LangChain Agents, or custom task schedulers to manage agent workflows, tool-use, and state persistence. Implement secure, auditable agent pipelines with human-in-the-loop checkpoints, logging, and governance controls. Containerize and deploy ML and agentic components with Kubernetes or serverless platforms; build CI/CD for model, RAG-component, KG, and agent updates. Monitor system performance end-to-end: model drift, retrieval quality, graph consistency, agent accuracy, and overall SLA/SLI adherence. Automate alerting, rollback strategies, and canary-based rollouts to minimize runtime risk. Stay current on advances in RAG architectures, graph-based neural models, autonomous agent research, and orchestration tooling; propose POCs or pilot projects.
experience
6 
