Kusum Sai Chowdary Sannapaneni
Machine Learning Engineer @Apple
Signup · Get unlimited contacts
WORK HISTORY
Machine Learning Engineer @Apple
CA, US
Designed and deployed federated learning models using TensorFlow Federated (TFF) and Core ML to enhance on-device personalization for Siri, ensuring data privacy and GDPR compliance.• Built a hybrid NLP pipeline combining Scikit-learn, TF-IDF, feature selection, and TinyBERT fine-tuning, improving multilingual intent recognition accuracy by 22% and reducing inference latency by 30%.• Engineered prompt-tuning strategies for on-device LLMs (TinyBERT) to boost contextual understanding and follow-up response accuracy, increasing task completion rates by 17% with no added latency.• Developed a feature store using Python (Dask, SQLAlchemy) and AWS Aurora PostgreSQL, cutting training data latency by 40% and Improved data traceability and quality by implementing lineage tracking and schema validation across training datasets.• Created a Neural Architecture Search (NAS) framework using evolutionary algorithms to optimize wake-word detection models, reducing false positives by 25% while maintaining sub-100ms latency on Apple M-series chips.• Led MLOps automation initiatives with Kubeflow, AWS SageMaker, and CI/CD pipelines, accelerating model deployment cycles from 2 weeks to 3 days.• Mentored junior engineers on optimization algorithms (Lagrange multipliers, gradient descent), advancing skills in model compression and on-device inference.
EDUCATION
The George Washington University
Master of Science - MS
ABOUT KUSUM SAI CHOWDARY SANNAPANENI
Machine Learning Engineer with 3+ years of experience designing and deploying production-grade AI/ML solutions at scale for top-tier tech and enterprise clients. Proven expertise in federated learning, NLP, LLM fine-tuning, time-series forecasting, and model optimization, with a strong foundation in MLOps, cloud platforms (AWS, Azure), and data engineering pipelines. Adept at building scalable ML systems using tools like TensorFlow, PyTorch, Scikit-learn, Airflow, PySpark, and SageMaker, and deploying models with CI/CD, Docker, and Kubernetes. Known for driving measurable improvements in accuracy, latency, and operational efficiency, while ensuring GDPR compliance and model governance.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.