Kermi Kotecha

I am a result-driven AIML Engineer with expertise in Python,ROS, AI and Machine Learning,Computer Vision and Generative AI.Developed innovative AI solutions, contributing to real-time object detection and advanced AI dashboards.I am highly motivated and dedicated with a strong work ethic, a passion for continuous learning, and a collaborative team-oriented mindset. I thrive on tackling challenges head-on and embrace opportunities to expand knowledge and skills. Excited to contribute expertise and enthusiasm to a dynamic and innovative environment.

LLM based Network Anamoly Detection (CDAC) 

Developed a random forest based GPT2 LLM to generate alerts for network anamoly detection. Gives the attack category as well as sub category along with preventative measures as output to the user in human understandable language

AIoT based sustainable health solution for wind turbine blades
Developed a Deep-learning based system that predicts the health of a wind turbine based on the vibrations produced by the wind blades
Aids in early fault detection and maintenance of the blades
Integrating a LLM to generate alerts when the blade needs attention thus preventing loss of resources

Multimodal Neuro physiological framework for Cognitive Behavior Analysis
Developed a multimodal neural network based on multiple modalities
such as EEG, ECG, eye tracking, Audio, Video, gaze and GSR to create a
model for stress and lie detection
collected real time data, developed deep learning models for the analysis and used late fusion for creating multimodal pipeline

Flat Bath Preddicition System (Aston University UK)
Developed a model to predict weather the furnance has reached its maximum temperature and the metal ore has metled based on the temperature of the furnance roof aids in preventing loss of heat and resources while melting iron ores

NLP based Finance Document Summarizer
developed a finance document summariser using NLP
preprocessed the data and applied FinBert model to summarise the
documents

Improving emotion detection in audio using deep learning algorithms
Used SAVEE dataset which consists of audio files
Preprocessed the audio files and generated spectrograms using MFCC
and librosa libraries
Applied machine learning and deep learning algorithms on the data to
classify the data.

Comparative Analysis of classification algorithms for Recommendation system
Using books dataset we compared different types of classification algorithms such as BIRCH, KNN,K-means

• Supervised Machine Learning: Regression and Classification, Coursera
• EDA for Machine Learning, Udemy
• C Programming Bootcamp - The Complete C Language Course, Udemy
• Learning Python for Data Analysis and Visualization, Udemy
• 2022 Complete Python Bootcamp From Zero To Hero in Python, Udemy
• Complete Python Image Processing Masterclass, Udemy
• Advance Learning Algorithms, Coursera
• Introduction to Cloud Computing, IBM
• Foundations of Data Science, Google

• SPICES grant from AICTE Along with our mentor Dr Neha Diver we applied for and won the SPICES (scheme for promoting interest, creativity and ethics among students) grant from AICTE worth Rs 1 Lakh
• SDG (sustainable development goals) Grant from United Arab Emirates University, Abu Dhabi working on a project in collaboration with UAEU titled AIoT based sustainable health solution for wind turbine blades.
• IEEE travel grant Received an IEEE grant worth 700 USD to attend summer school on the topic Envisioning the future- Pioneering Next Generation Communication Technologies in Sri Lanka