Personal details
Akhil M.
Technical skills
About
Data Scientist with 5 years of experience in developing AI/ML models, NLP applications, and data-driven solutions to meet business requirements. Expertise in designing predictive models, deploying machine learning systems, and improving model accuracy using Python, PySpark, and Azure. Proven ability to enhance process efficiency and customer satisfaction through innovative AI solutions and data analytics. Strong skills in Power BI, MS Azure, and deployment in production environments.
Work Experience
- Collaborated closely with cross-functional teams to conduct thorough analysis of extensive sales data, resulting in a substantial improvement in prediction accuracy.
- Spearheaded the training of a Deep Neural Network within the Keras framework to forecast a 2-hour delivery window for a logistics client's package delivery. This involved comprehensive data pre-processing and training on Azure Databricks. Subsequent post-modeling analysis, including the creation of a confusion matrix and computation of accuracies, was carried out for an ensemble model.
- Led the end-to-end design, development, and seamless implementation of advanced ML Pipelines, which significantly enhanced the prediction capabilities for sales failures.
- Engineered and deployed a sophisticated web application for document question-answering, leveraging the GPT4ALL Large Language Model (LLM) and LangChain framework. The application is poised for deployment on Streamlit to ensure user-friendly access.
- Leveraged Microsoft Azure's state-of-the-art services to drive efficiency in both application development and deployment processes, particularly in the realm of MLOps.
- Demonstrated expertise in managing the application development lifecycle using Microsoft Azure DevOps, ensuring a smooth and efficient workflow.
- Exemplified strong leadership and mentorship qualities by guiding a team of 6 junior associates in developing expertise in applications with GPT4ALL LLM and LangChain framework.
• Developing an ongoing machine learning-based system for detecting and managing crop diseases in maize and tomato crops in Tanzania.
• Collaborating with a multidisciplinary team on image analysis techniques and CNNs for accurate disease detection.
• Collecting a comprehensive dataset of healthy and diseased crop images, while contributing to user-friendly interface development.