Anna Ullah

Designed and deployed LLM and Transformer-based NLP models using Hugging Face, BERT, and RoBERTa. Developed cloud-native ML APIs using FastAPI, containerized with Docker, and deployed to AWS Lambda. Implemented real-time inference monitoring with Grafana and Prometheus to track model health. Built automated model retraining pipelines using MLflow and GitHub Actions with data versioning. Worked with software engineers to integrate model outputs into backend systems via RESTful services. Supported model governance by maintaining artifacts, lineage, and reproducibility standards. Led technical documentation and onboarding for new MLOps team members.

Designed and deployed LLM and Transformer-based NLP models using Hugging Face, BERT, and RoBERTa. Developed cloud-native ML APIs using FastAPI, containerized with Docker, and deployed to AWS Lambda. Implemented real-time inference monitoring with Grafana and Prometheus to track model health. Built automated model retraining pipelines using MLflow and GitHub Actions with data versioning. Worked with software engineers to integrate model outputs into backend systems via RESTful services. Supported model governance by maintaining artifacts, lineage, and reproducibility standards. Led technical documentation and onboarding for new MLOps team members.

Developed recommendation systems and classification models using PyTorch and Scikit-learn. Managed infrastructure for distributed model training across AWS and GCP with autoscaling enabled. Built data ingestion pipelines using Apache Airflow and optimized query performance using BigQuery. Designed modular codebase for experimentation and testing with MLflow tracking. Collaborated with frontend and backend teams to integrate ML services into customer-facing applications. Conducted weekly code reviews and mentored junior ML engineers on model deployment best practices. Applied advanced feature engineering techniques using time-series and categorical data.