Sufyan Zulfiqar

About

Production-focused ML/AI engineer with hands-on ownership across classical ML and GenAI systems, from CV/NLP services to enterprise RAG. Stands out for combining research-to-production execution with measurable business impact: 40% processing-efficiency gains, 35% fewer support tickets, 5x latency improvement, and 3x throughput gains while maintaining safety and quality.

Key Strengths

• Owned end-to-end production ML systems from R&D through deployment and monitoring
• Improved CV/NLP processing efficiency by 40%
• Strong MLOps and scalability experience with Kubernetes, CI/CD, and drift monitoring
• Optimized inference latency using ONNX and TensorRT
• Shipped enterprise RAG system with hybrid search and grounded retrieval
• Reduced support ticket volume by 35% with 95% response accuracy
• Applied rigorous LLM evaluation and safety methods using RAGAS, G-Eval, toxicity controls, and red teaming
• Handled fine-tuning tradeoffs with PEFT/LoRA to mitigate catastrophic forgetting
• Translated knowledge distillation research into production with 5x latency improvement and lower compute cost
• Boosted throughput 3x via dynamic batching without added infrastructure overhead
• Improved engineering velocity by reducing deployment cycle time 40%
• Increased reliability by eliminating 90% of recurring integration bugs

Experience

Education

Skills

Languages