dheeraj krovvidi

India

@dheerajpranav

AI Club | ML Engineer

Badges

Problem Solving
Python
Days of Code
Days ofStatistics

Certifications

Work Experience

  • AI Engineer

    AspireNxt•  May 2022 - Present

    - Building Pytorch models for Image Classification tasks & working on deployments, end-points using AWS. - Worked on building one-click deployment frameworks in programmatical way for: 1. Recommender systems using Python, Terraform on AWS Personalize. 2. Time Series forecasting for sales using AWS forecast 3. Anomaly Detection using Amazon L4M. * Tools / Tech Stack * : Python, Terraform, AWS services- Sage Maker, Amazon Lex, Personalize, Lambda Functions, Amazon Forecast & Anomaly Detection

  • Machine Learning Engineer

    Applicate•  April 2020 - April 2022

    - Driving projects from an architecture & framework point of view in the Machine Learning Team. - ML Framework : Worked on building the Generic machine learning frameworks for highly scalable and distribution ETL jobs with automatic re-processing and 100% automations. Used the end to end pipelines, jobs & automated systems in AI for deploying simplified solutions at scale. Serving them as a single platform in Python 3. - Sales Forecast : Designed a customized Statistical Quantile approach to forecast sales/quantity which increased model performance by 20% along with time optimization. Improved overall performances with better design choices & asynchronous behaviour enforcing state less machines and idempotent algorithms. - Recommendation Systems : Building Personalized Recommendation systems for suggestions of sales products to end-users. Used Hierarchical clustering & Ranking methods for Similarity & Relevance score generation. Used advanced parallelization method Dask to process, handle large data sets & to reduce data retrieval time by 75%. - Explainable AI (EDA) : Developed EDA Dashboards on Kibana to monitor and manage Insights of data accuracy metrics for compliance reports on regular scale. Built a customized error-metric to categorize Valid & Inconsistent data samples based on purchase pattern. For Valid cases we've achieved accuracy > 85 % and more precise for most of the clients. - Tools : Python, CI/CD pipelines (ETL jobs), Kibana, MySQL, Jenkins, AWS-S3, Machine Learning Algorithms - XGBoost , Random Forest, Other Classification & Clustering Approaches.

  • ML Intern

    Inversion consultancy LLP•  May 2019 - August 2020

    I was exposed to the various activities in the Machine Learning and Deep Learning domains using Python and TensorFlow, implementing Neural Networks, Predictive models, Regression, Recommendation Engines, and many more skills.

Education

  • Vignan Institute of Technology and Science, Yadadri Bhuvanagiri

    Computer Science & Engineering, B.Tech•  August 2016 - May 2020

Skills

dheerajpranav has not updated skills details yet.