Machine Learning Ai Engineer

Lokesh Madem

******@***********.***

Innovative and results-driven AI Engineer with 3+ years of specialized experience in developing and deploying advanced machine learning models, computer vision and NLP solutions in high-stakes environments. Adept at conducting rigorous data processing, model development, and performance evaluation, using a data-driven approach to achieveand surpass project goals. Committed to continuous learning and application of the latest AI advancements to drive business value and innovation

+1-940-***-**** Denton,TX LinkedIn

AI/ML Engineer

TECHNICAL SKILLS

Programming Languages: Python, SQL, HTML, CSS, JavaScript, Java, R, MATLAB, C++, C Artificial Intelligence & Machine Learning Frameworks: TensorFlow, PyTorch, Keras, Hugging Face, Langchain, FinBERT, Faster R - CNN, YOLO, LSTM, CNN, ResNet50, K-Means Clustering, MediaPipe, mAP, F1 score, Bleu, ROGUE, METEOR, BERT, Numpy, Pandas, Matplotlib

NLP Tools: NLTK, spaCy, Transformer Models, Llama, Mistral, GPT Cloud Platforms & Services: AWS (Lambda, SageMaker, Glue, ECS, EC2, Eventbridge, StepFunctions, Cloudwatch) Tools & Technology: Pinecone, Flask, FastAPI, Docker, Kubernetes, Prometheus, Grafana, MySQL, Spark,DataBricks, DynamoDB, Airflow, Power BI, Tableau, JIRA, Jenkins, Git, GitHub EXPERIENCE

Spearheaded the collection and preprocessing of financial data using Python, NLTK, and spaCy, ensuring compliance with data privacy standards while processing transaction details, customer interactions, and payment notes. Built advanced NLP pipelines featuring tokenization, lemmatization, and data vectorization, leveraging TensorFlow and Hugging Face models to efficiently process and analyze large financial datasets. Designed and fine-tuned a FinBERT-based model for classifying financial transactions into categories like groceries, subscriptions, and travel, boosting classification accuracy by 20% and providing personalized financial insights. Developed a Retrieval-Augmented Generation (RAG) chatbot using GPT, Llama, Mistral, and Pinecone to deliver context- aware, real-time customer support with improved relevance and accuracy of responses. Orchestrated model deployment by integrating AWS Lambda, SageMaker, and ECS for scalable real-time inference of transaction classifications and AI-powered chatbot interactions. Implemented a real-time monitoring system with Prometheus and Grafana to track key metrics—accuracy, precision, recall, and F1-score—ensuring consistent performance and minimizing data inconsistencies. AI Engineer, PayPal September 2023 – Present Remote,USA AI Engineer, L & T Technology Services April 2021 – Jul 2022 Bangalore, India Developed object detection models, including Faster R-CNN and YOLO using PyTorch, to detect and recognize worker safety equipment such as helmets, reducing safety incidents by over 25% and increasing workplace compliance monitoring speed by 40%.

Managed the collection and annotation of over 10,000 images of workers in industrial environments, collaborating with domain experts to train and validate AI models.

Implemented data preprocessing and augmentation techniques that resulted in a 20% improvement in model robustness and performance.

Set up automated testing frameworks to continuously track the system’s mAP and other relevant metrics such as precision, recall, and F1-score, enabling quick identification of issues during development cycles. Programmed backend services in Python, incorporating robust error logging and performance monitoring tools that ensured an operational uptime of 99%. Deployed services in containerized environments using Docker for portability and consistency.

Developed and deployed cloud-based AI-powered safety monitoring services using AWS, orchestrated with Kubernetes, enabling scalable deployment of object detection models for real-time recognition of worker safety equipment compliance via RESTful APIs, enhancing workplace safety and operational efficiency. University of North Texas Aug 2022 – May 2024

Master of Science; Major in Artificial Intelligence Denton, USA Andhra University Jul 2018 – May 2022

Bachelor of Technology; Major in Computer Science Andhra Pradesh, India EDUCATION

Summary

Neural Network & Deep Learning – Coursera

Improving Deep Neural Network: Hyperparameter Tuning, Regularization & Optimization - Coursera Python for Everybody Specialization – Coursera

CERTIFICATIONS

LanguageTranslator

Technology Used : NLP, AWS, PyTorch, MLOps, Transformer, Docker Developed a real-time language translation model using transformer architecture in PyTorch, trained on a dataset of 0.45 million English-Telugu sentence pairs, achieving a BLEU score of 0.55. Deployed an end-to-end MLaaS framework for English-to-Telugu translation, encompassing data acquisition, preparation, model tuning, and API implementation with Flask and Postman. Successfully hosted the model on the Waisum platform (https://www.waisum.co/), a collaborative MLaaS environment supporting blog hosting and ML model prototyping with a capacity for 1000+ concurrent users. Automatic Image Captioner

Technology used: ComputerVision, Natural Language Processing, TensorFlow, NLTK, CNN, LSTM Successfully developed a model that integrates CNN for feature extraction from images and LSTM for generating coherent and contextually appropriate captions.

Achieved a high BLEU score of 0.663, which validates the model's proficiency in synthesizing accurate and relevant captions, thus demonstrating advanced capabilities in linking visual data with natural language processing. Demonstrated the system’s potential to enhance accessibility features in digital media platforms by providing automatic text descriptions, facilitating a better understanding of visual content for all users, including those with visual impairments. Automatic Gym Trainer

Technology used: Computer Vision, MediaPipe, Flask, HTML, CSS, KMeans Clustering, Pattern Recognition Engineered a robust solution that utilizes MediaPipe for real-time extraction of key points, enabling accurate tracking of human body movements.

Implemented KMeans clustering to identify and categorize exercise patterns effectively, achieving high accuracy in exercise recognition and counting, thus ensuring system reliability. Developed a user-friendly web interface using Flask, HTML, and CSS, which allows users to interact with the system easily through a web browser. Also provided API access for potential integration with other applications, enhancing the system’s usability and accessibility.

Heart Disease Prediction System

Technology used: Flask, Random Forest Classifier, SQL, HTML, CSS Developed a RandomForest model that achieved a high prediction accuracy of 90.16%,meeting rigorous performance standards and demonstrating the ability to handle and analyze large datasets effectively. Optimized model performance throughmeticulous data analysis and tuning, ensuring robust and reliable predictions. Implemented a seamless data management system using an SQL database, which supports efficient handling of large data volumes and underpins the API for easy integration. Developed a user-friendly Flask web application, providing a real-time prediction interface that allows users to receive immediate health insights through a simple web-based platform, enhancing user engagement and accessibility. ACADEMIC PROJECT

Contact this candidate