ML Engineer focused on production GenAI, LLM applications, and MLOps.

I build reliable AI systems that move from experimentation to real-world deployment, with emphasis on retrieval quality, agent reliability, and observable ML infrastructure.

• Build and deploy end-to-end LLM systems: ingestion, retrieval, orchestration, serving, monitoring
• Design RAG pipelines with hybrid retrieval, reranking, and evaluation-driven optimization
• Develop agentic workflows with tool use, memory, planning, and guardrails
• Productionize ML with CI/CD, containerized deployment, and cloud-native operations
• Apply strong engineering standards: testing, observability, reproducibility, and security

• Prompt engineering, model adaptation, and response-quality optimization
• RAG architecture design with chunking, indexing, reranking, and query routing
• Agent frameworks for multi-step task execution and tool orchestration
• LLM evaluation pipelines for quality, latency, and cost trade-offs
• Multi-modal pipelines for text, image, and audio use cases

• NLP: Transformers, embedding systems, semantic search, information extraction
• Computer vision: detection, segmentation, and classification workflows
• Model lifecycle: training, validation, packaging, and deployment
• Inference optimization: batching, quantization, caching, and distillation

• Containerized deployment with autoscaling and service reliability practices
• CI/CD for ML and backend systems with automated checks and rollout safety
• Observability: logs, metrics, traces, and model/data drift monitoring
• Infrastructure as code with reproducible environments and versioned pipelines
• Data and experiment management for auditable model development

• Multi-agent orchestration with dynamic tool selection and task planning
• Context-aware agents with conversation memory and stateful workflows
• Tool-augmented LLM systems for automation, analytics, and developer productivity

• End-to-end pipelines for ingestion, indexing, retrieval, reranking, and response synthesis
• Hybrid retrieval stacks combining dense and sparse strategies
• Evaluation loops to improve answer quality, grounding, and hallucination resistance

• Real-time inference APIs for latency-sensitive workloads
• Batch prediction pipelines for high-volume offline processing
• Vision and NLP applications deployed as reliable services

• Automated training, validation, and release workflows
• Model and data monitoring with drift and performance tracking
• Scalable serving systems with reliability, rollback, and cost controls

• Reliability: strong testing, robust error handling, graceful degradation
• Scalability: horizontal scaling, caching, efficient resource utilization
• Observability: actionable metrics, structured logging, distributed tracing
• Reproducibility: versioned code/data/models and deterministic workflows
• Security: authentication, rate limiting, and secure secret management
• Cost awareness: right-sized infrastructure and efficient inference patterns

• Email: [email protected]
• Email: [email protected]
• LinkedIn: linkedin.com/in/reeth-jain-rj777

Building AI systems that scale from prototype to production.