Experience — Muhammad Nouman Khan

AUG 2025 — PRESENT CURRENT FULL-TIME

MLOps Engineer

BetaCodes Pvt Ltd 📍 Islamabad, Pakistan

Provisioned and managed NVIDIA DGX systems and GPU nodes (H100, H200) for distributed AI model training and high-throughput inference workloads.
Managed Kubernetes clusters using Kubeadm, Terraform, and BCM (Base Command Manager) for high-availability production environments.
Built AI model serving infrastructure using vLLM and Triton Inference Server; applied quantization (INT4/INT8), tensor parallelism, and continuous batching to maximize GPU utilization.
Implemented model scaling strategies including horizontal pod autoscaling and GPU-aware K8s scheduling to handle variable production loads.
Designed end-to-end ML pipelines with CI/CD, Docker, and cloud-native workflows on AWS, GCP, and Azure — from model development to production.
Deployed comprehensive monitoring, alerting, and observability stacks for production AI models ensuring SLA compliance.

NVIDIA DGX H100/H200 Kubernetes Kubeadm Terraform BCM vLLM Triton Kserve Docker CI/CD AWS GCP

AUG 2024 — PRESENT CURRENT REMOTE · USA

AI Engineer

iQera Schools 🌐 Remote (USA)

Developed and deployed Retrieval-Augmented Generation (RAG) systems and agentic AI frameworks, enhancing the iQera Schools e-learning platform with generative AI capabilities.
Built scalable backend services with FastAPI and Django, integrating vector databases for semantic search and comprehensive educational resource access.
Managed cloud infrastructure on AWS including model hosting, API gateways, and vector database deployments.
Led project management for AI feature delivery across cross-functional teams spanning the US and Pakistan.

RAG Agentic AI FastAPI Django AWS Vector DBs LangChain LlamaIndex

JUL 2024 — SEP 2024 INTERNSHIP

GenAI Intern

Sybrid Pvt Ltd 📍 Islamabad, Pakistan

Fine-tuned Stable Diffusion models (v1.5, 2.1, 3, and Flux1-dev) for generative AI tasks by building and captioning a localized demographic dataset.
Enhanced model performance on clothing, facial imagery, architecture, and natural scenes using PyTorch and HuggingFace Transformers.
Designed and ran systematic fine-tuning experiments with DreamBooth and LoRA techniques for targeted domain adaptation.

Stable Diffusion Flux1-dev LoRA DreamBooth PyTorch HuggingFace Generative AI

AUG 2023 — SEP 2023 INTERNSHIP

ML / Computer Vision Intern

Rapidev 📍 Islamabad, Pakistan

Built a production-grade car number plate detection model using YOLO, deployed via RESTful APIs for real-time vehicle identification.
Developed APIs for raster-to-vector conversion used in engineering document digitization workflows.
Annotated and curated training datasets using CVAT and Roboflow ensuring high data quality for production deployments.

YOLO OpenCV CVAT Roboflow REST APIs Python

JUL 2022 — SEP 2022 INTERNSHIP

AI Intern

Sino-Pak Center for Artificial Intelligence (SPCAI) 📍 PAF IAST, Pakistan

Deployed machine learning models on edge devices using Raspberry Pi and OpenCV OAK-D for real-time inference in resource-constrained environments.
Built a facial attendance management system with live face recognition and database integration for automated record keeping.
Developed a solar irradiance forecasting model for smart building energy optimization using time-series ML techniques.

Edge AI Raspberry Pi OpenCV OAK-D Face Recognition Time-Series Forecasting

Career Timeline