Avinash Nagabelly
I am an AI/ML engineer with over 4 years of experience designing, building, and deploying AI solutions across healthcare, life sciences, and enterprise domains.
About
I am an AI/ML engineer with over 4 years of experience designing, building, and deploying AI solutions across healthcare, life sciences, and enterprise domains. I specialize in deep learning, NLP, computer vision, predictive modeling, and generative AI, and I routinely architect end-to-end data pipelines, MLOps workflows, and cloud-based deployments. I thrive on translating complex business and clinical requirements into reliable, high-impact AI/ML solutions.
I excel at feature engineering, model optimization, interpretability, and performance monitoring, and I collaborate closely with cross-functional teams to deliver measurable outcomes. I am proficient in deploying scalable inference services, maintaining reproducible experimentation, and driving continuous improvement through robust monitoring and retraining pipelines.
- 5 years' experience
- Joined April 2026
Industry experience
- Healthcare
- Life Sciences
- Professional Services
- Software & Internet
Work experience· 4 roles
Developed oncology-focused survival prediction models using XGBoost and regression on structured EHR data; implemented GPT-4–driven clinical NLP pipelines in a LangChain-based RAG setup to extract structured oncology variables from unstructured notes; evaluated BERT-based oncology entity extraction and benchmarked summarization against Claude; applied SHAP/LIME for interpretability; managed experiments with MLflow and Model Registry; automated retraining with Apache Airflow; deployed Docker-based inference services on AWS SageMaker for batch scoring of oncology datasets (~20k records/month).
Developed Time Series Forecasting models using scikit-learn to analyze clinical trial enrollment variability; performed distributed feature engineering on high-volume life sciences datasets with Apache Spark on Azure Databricks; built compliance-focused Text Classification models using logistic regression on MySQL datasets (~120,000 records).
Architected XGBoost and CatBoost regression and clustering models for KPI forecasting, improving prediction accuracy by 22% and reducing forecast variance by 18% across portfolio projections. Automated document classification pipelines using NER deployed on Azure Databricks, with outputs stored in MySQL, reducing misclassification errors by 28% in regulatory workflows. Integrated CI/CD with GitHub Actions, Docker, and Kubernetes on Azure ML, reducing failures by 12%, shortening release cycles by 35%, and saving 6,000+ manual deployment hours annually. Monitored ML pipelines on a shared MLOps platform with Datadog and Azure Monitor, sustaining 99.5% uptime and triaging 3,000+ alerts annually. Implemented Reinforcement Learning models in simulated decision environments to evaluate policy trade-offs, reducing forecasting error by 14% across 1,000+ scenarios. Delivered Power BI dashboards for leadership visibility, reducing decision latency.
Education
Certifications and awards
Hire Avinash Nagabelly today
Post a job and invite Avinash, or contact them directly. No hidden commission.






