S
Sachin Miryala
About
Detail
Gujarat, India
Résumé
Jobs
verified_user
0% verified
AI/ML EngineerKeyBankMar 2024 - Current (2 years 8 months)• Worked on building and training advanced AI models using BERT and deep learning for customer segmentation and personalization. Collaborated with stakeholders and cross-functional teams during requirement gathering sessions, boosting targeted campaign effectiveness by 20%. • Developed AI models and engineered features using Python, Pandas, and NumPy, processing financial datasets including transaction history, customer feedback, and demographics stored in AWS Redshift and DynamoDB. Applied feature selection and dimensionality reduction to improve model inputs. • Trained models with TensorFlow, PyTorch, and scikit-learn, leveraging large-scale AWS-based SQL and NoSQL databases for efficient data ingestion and preprocessing pipelines, sig- AI/ML EngineerNovartisJan 2019 - Dec 2022 (4 years)• Built and trained predictive AI models including XGBoost and Random Forest using Python, improving drug efficacy prediction accuracy by 18% through advanced feature engineering from clinical and biomedical datasets, enhancing decision-making in drug discovery. • Applied NLP models using SpaCy and NLTK to extract features from unstructured clinical texts, increasing dataset richness by 25%, which improved side effect prediction precision by 15% and overall model robustness. • Conducted hyperparameter tuning of XGBoost and Random Forest models using GridSearchCV and Optuna, boosting model recall and precision by 20%, while decreasing training time by 30% for scalable, efficient ML pipelines. • Tested models rigorously through A/B testin
- ML EngineerHSBCJan 2018 - Dec 2018 (1 year)• Collaborated closely with senior data scientists and business analysts to craft the customer segmentation strategy using clustering algorithms, improving targeting accuracy by 30% through detailed analysis of HSBC customer banking behavior and preferences. • Led data preprocessing and feature engineering using Python (Pandas, NumPy), increasing data quality and model input relevance by 25% through cleaning, transformation, and applying PCA for dimensionality reduction to enhance model efficiency. • Built and trained clustering models such as K-means and DBSCAN with Scikit-learn, tuning hyperparameters via grid search that improved cluster coherence by 20%, and created visualizations with Matplotlib and Seaborn for effective stakeholder
Education
verified_user
0% verified
- Master of ScienceNorthern Arizona UniversityJan 2023 - Dec 2024 (2 years)
- Bachelor of EngineeringPanimalar Engineering CollegeAug 2014 - May 2018 (3 years 10 months)