P
Pavani Mengaraboina
About
Detail
Gujarat, India
Résumé
Jobs
verified_user
0% verified
AI/ML EngineerBrexJan 2024 - Current (2 years 6 months)• Worked on developing transformer-based NLP models to accurately extract, classify, and validate financial data from invoices, receipts, and statements. Reduced manual processing time by 60 percent and improved compliance through close collaboration with product and finance teams during requirement gathering sessions. • Built and deployed scalable cloud-native data pipelines using AWS Textract and SageMaker. Enabled real-time processing of millions of transactions each month while integrating seamlessly with Brex's core financial systems. • Designed and implemented anomaly detection and ensemble learning models to identify fraudulent activity in real time. Increased fraud detection accuracy by 35 percent while minimizing false positives- AI/ML EngineerMichael PageJun 2021 - Dec 2022 (1 year 7 months)• Collaborated with HR stakeholders and data scientists to gather requirements and design transformer-based NLP models using Python and Azure Cognitive Services, enhancing candidate sentiment analysis and improving recruitment decision-making. • Directed end-to-end model development, building automated ML pipelines with Python libraries like Hugging Face and Scikit-learn, boosting accuracy of job market trend predictions by 30%, and documented workflows for seamless team adoption. • Extracted and preprocessed candidate and job data using SQL, ensuring high-quality datasets for Azure ML workflows, enabling reliable and scalable talent analytics and workforce planning. • Implemented distributed training on Azure GPU clusters, reducing tra
- FMachine Learning EngineerFranklin Templeton IndiaJun 2017 - Aug 2019 (2 years 3 months)• Designed and implemented a RL framework using Markov Decision Processes (MDP) to address complex portfolio allocation challenges, collaborating closely with finance experts to develop reward functions that balanced risk and return, boosting decision accuracy by 22%. • Developed the RL training environment and data processing pipeline using Python with key libraries such as NumPy, Pandas, and TensorFlow 1.x, while using SQL to efficiently extract and preprocess large-scale historical financial datasets. • Created realistic market scenario simulations for training the RL model, ensuring robustness by incorporating diverse historical market conditions and data variations. • Executed wide backtesting of RL policies using financial metrics
Education
verified_user
0% verified
- Master of ScienceUniversity of Central MissouriJan 2023 - May 2024 (1 year 5 months)
- Master of ScienceOsmania UniversitySep 2019 - Jul 2021 (1 year 11 months)
- Bachelor of ScienceOsmania UniversityJun 2014 - May 2017 (3 years)