I am Medhaja H L

Generative AI Engineer,MLOps Engineer,ML Engineer,Tech Savvy

Name: Medhaja H L

Profile: Senior Software Engineer (Gen AI/ MLOps)

Email: medhajahlbhat@gmail.com

Skill

Python 85%
C 75%
Java 50%
About me

I am a Graduate in Computer Science & Engineering from Manipal Institute of Technology, Manipal.

I have experience of development and deployment of various Gen AI and ML models based on RAG, Chat bots, CNN, LSTM, Auto Encoders, YOLO , Differential Privacy etc.

The incredible rush of making things happen inspires me to push boundaries. Open for full-time roles such as Generative AI Engineer, Software Engineer, Machine Learning Engineer, Data Scientist, MLOps Engineer and aligning roles. Open to relocate Internationally.

Projects

AADHAAR CARD DATA EXTRACTION FOR KYC WITH YOLO AND AI-OCR

Extracting Data from Aadhaar card using YoloV5. Accuracy of 95% was achieved. Datasets contained 35 images in the training set, 5 images in validation set & 5 images in the test set. Employed Data Augmentation and Expanded the Data set by 40x. The training as well as validation accuracy was 86% and 83% considering the small amount of data set.

Heart Failure Prediction using XGBoost and MLflow

Demo Link

Developed a heart failure prediction model using XGBoost algorithm on a clinical dataset with 13 features and 300+ instances, achieving high accuracy and ROC AUC scores. Utilized MLflow for experiment tracking, model management, and hyperparameter tuning to improve model performance. Visualized feature importance to gain insights into the impact of different clinical features on heart failure prediction.

COLA-POWER BAR CLASSIFICATION USING VGG16 MODEL(TRANSFER LEARNING)

Demo Link.

Binary classification of Coco cola and Power bar photo. After fine tuning the VGG16 model to fit to our Coco-cola and Power bar dataset, the training accuracy went upto 95.6% & Validation Accuracy was 92.7%

HAND WRITTEN CHARACTER RECOGNITION BASED ON VISUAL INPUT USING NEURAL NETWORK

This is a research project in which we can Recognizes English Alphabets by the movement of the Pen's cap.

APPLICATION OF DIFFERENTIAL PRIVACY FOR TRANSFER LEARNING

Implementing Differential Privacy for Transfer Learning model using VGG16 architecture.

APPLICATION OF DIFFERENTIAL PRIVACY FOR BINARY CLASSIFICATION

Implementing Differential Privacy for Binary Image Classification using CNN architecture.

A MACHINE LEARNING APPROACH TO PRODUCT REVIEW DISAMBIGUATION BASED ON FUNCTION, FORM AND BEHAVIOR

This is a research project in which we Classify a Product reviews into one of the five categories: Form, Function, Behavior, Services, Other.

6

PROJECTS COMPLETED

5

YEARS OF PROGRAMMING EXPERIENCE

3.4

TOTAL INDUSTRY EXPERIENCE

1

RESEARCH PAPER

PROFESSIONAL DEVELOPMENT & AF­FILIATIONS

Technical certifications earned in various online platforms

Sustainable Business Strategy

HBS Online | HBS

Large Language Model Operations Specialization

Coursera | Duke University

AI Product Management Specialization

Coursera | Duke University

Machine Learning Engineering for Production (MLOps) Specialization

Coursera | Deep Learning.ai

Machine Learning Specialization

Coursera | University of Washington

Deep Learning Specialization

Coursera | Deep Learning.ai

Algorithms Specialization

Coursera | Stanford University

Applied Data Science Specialization

Coursera | University of Michigan

Deep Learning with Tensorflow

Cognitive Class

Big Data Modeling and Management Systems

Coursera | UC San Diego

Big Data Integration and Processing

Coursera | UC San Diego
Send Message Us
Your message has been sent. Thank you!
Get in Touch

I would like to get in touch

  • Bengaluru, Karnataka, India
  • medhajahlbhat@gmail.com