Hanif Lashari

About

Built and productionized an agentic RAG assistant using Ollama + LangChain + MCP + ChromaDB to speed up and standardize access to operational knowledge from tickets and runbooks. Focused on real-world reliability: mitigated timeouts/latency with retries and concurrency limits, improved retrieval via chunking/embedding iteration, and reduced hallucinations through citation-grounding and confidence-based abstention. Also partnered with non-technical ops staff to deliver anomaly detection/monitoring by translating operational needs into model signals, thresholds, and alerting logic.

Key Strengths

• Built and deployed an agentic RAG assistant over operational tickets/runbooks
• Stabilized LLM pipeline reliability via timeouts/retries and concurrency controls
• Improved retrieval quality through iterative chunking/embedding tuning and parameter tuning (e.g., top-K)
• Reduced hallucinations with strict grounding (citations), confidence thresholds, and safe abstention behavior
• Designed workflows as testable components (ingestion, retrieval, validation, generation) with offline test sets
• Measures and improves production performance using latency, retrieval relevance, and answerability metrics with step-level logging
• Effective collaboration with non-technical operational stakeholders using plain language and simple visualizations

Experience

Education

Skills

Languages