Aakash Dhakse
Turning Ideas into Digital Reality
Turning Ideas into Digital Reality
I am an AI Engineer with hands-on experience in building intelligent and scalable AI solutions using Python, FastAPI, and Django.
I specialize in developing AI-powered systems including image-based pet disease detection using YOLOv8 and OpenCV,
audio-based pet emotion recognition using MFCC and Wav2Vec, and computer vision pipelines for object and safety detection.
I have strong expertise in Generative AI technologies such as LangChain, LangGraph, HuggingFace, vector databases (Pinecone),
and modern LLM concepts including prompt engineering, embeddings, RAG systems, and model fine-tuning.
Alongside technical development, I actively focus on improving real-world deployment, system scalability,
and professional growth through continuous learning and project-driven innovation.
Working as an AI Engineer, I design and develop intelligent AI-driven systems focused on computer vision, audio processing, and Generative AI. I built AI-based pet disease detection solutions using YOLOv8 and OpenCV for image-based diagnosis support. I also developed pet emotion recognition models for dogs and cats using audio signals, applying MFCC feature extraction and fine-tuned Wav2Vec models for accurate emotion classification.
In addition, I worked on pet toy safety detection systems using object detection pipelines to identify unsafe or harmful toys. I integrated GPT-4 for intelligent analysis, explanation generation, and AI-assisted decision support. My responsibilities also included data annotation, dataset cleaning, preprocessing, and generating embeddings to improve model accuracy, robustness, and inference efficiency.
During my training at Indvibe Infotech, I worked on building an image detection system using computer vision and deep learning techniques. I designed and trained custom object detection models to identify and classify objects from images and live camera feeds. The system involved data collection, annotation, preprocessing, and model training to achieve reliable detection accuracy.
I implemented the complete image detection pipeline, including image input handling, real-time inference, and result visualization with bounding boxes and confidence scores. The project strengthened my understanding of model training workflows, performance optimization, and practical deployment considerations. Through this training, I gained hands-on experience in developing end-to-end computer vision solutions and applying AI concepts to real-world use cases.
During my internship at Interflow, I worked on building simple yet practical projects using C and C++ to strengthen my core programming and problem-solving skills. I developed console-based applications focusing on logic building, data handling, and efficient use of control structures, functions, and arrays.
One of the key projects I built was a Student Management System using C++, where users could add, update, delete, and view student records through a menu-driven interface. The project used file handling to store data persistently and applied object-oriented concepts such as classes and encapsulation. This internship helped me gain a strong foundation in structured programming, memory management, and writing clean, efficient code.
HTML, CSS, JavaScript
Python, FastAPI, Django, C, C++
LangChain, LangGraph, HuggingFace, vector Databse(Pinecone)
Prompt engineering, embeddings, RAG, fine-tuning models. Github, Docker