Arshaq Kirmani

AWS Certified Machine Learning Engineer Associate with experience building multi-agent systems, cloud-based RAG applications, and end-to-end ML pipelines, driven to turn AI and machine learning into solutions for high-impact problems.

AWS AI Intern at Cloudelligent (2026-07 – 2026-10)

• Developed a RAG chatbot on Amazon Bedrock Knowledge Bases with 5/5 correct retrievals and 5/5 correct out-of-scope refusals; hierarchical chunking outperformed fixed-size chunking by resolving over-refusals.
• Built a customer support multi-agent system on Bedrock AgentCore using the Strands framework, with supervisor routing, an S3 Vectors knowledge base, and Guardrails for PII masking and denied topics.
• Trained and tuned an XGBoost churn model on Amazon SageMaker with hyperparameter optimization, achieving a ROC-AUC of 0.85, and designed an MLOps pipeline with Model Registry, CI/CD and drift-triggered retraining.

AI Intern at National Center of Artificial Intelligence (2025-07 – 2025-09)

• Integrated an LLM using Ollama, applied prompt engineering for clinical report updates, and optimized inference for a multimodal medical system handling patient intake, reporting, and transcription.

Teaching Assistant — Generative AI at FAST NUCES (2026-01 – 2026-05)

• Conducted demos and graded assessments for a Generative AI course covering LLMs and agentic AI concepts.

Bachelor of Science in Computer Science – FAST National University (2022-08 – 2026-06)