Anzar A.

Background

About Anzar A.

AI should be as reliable as electricity that’s the standard I build for. I design and deliver LLM-powered systems that scale bridging GenAI, MLOps, and cloud orchestration to move AI from prototype to production. AI & ML systems I’ve built: Retrieval-Augmented Generation (RAG): Cut research and ops time by ~70% with hybrid + semantic search pipelines. LangChain + LLM Orchestration: Automated compliance and reporting workflows, removing hours of manual work. Helm + Kubernetes Deployments: Ensured reproducible, observable, and rollback-ready ML/LLM workloads. Vector Search Optimisation: Improved enterprise search accuracy by 35% with hybrid retrieval + reranking. High-Scale APIs: Designed services processing 10k+ daily requests with >99.9% uptime on Azure/AWS. Core Toolbox: LLMs & GenAI: RAG • LangChain • embeddings • rerankers • agent frameworks • multimodal AI MLOps: Helm charts • CI/CD for ML/LLM apps • model registries • eval harnesses • monitoring • rollback plans Machine Learning: Python (pandas, scikit-learn, PyTorch) • data pipelines • feature engineering Cloud & Ops: Azure • AWS • Kubernetes • Docker • Terraform Backend & Delivery: FastAPI • Java (Spring Boot) • microservices • REST APIs Automation & Visualisation: Power Automate • Azure Functions • React dashboards • Power BI What sets me apart: I don’t just experiment with AI — I deliver enterprise-grade systems that are fast, resilient, and trusted in production. Future focus: multimodal agents, cost-aware LLM deployment, and Helm-driven AI orchestration at scale. Exploring LLMs, MLOps, or AI adoption at scale? Let’s connect.