Open to opportunities

Aditi Kohli

@aditikohli

Message

Aspiring AI/ML engineer using Python, deep learning, and web development to build impactful, accurate models.

India

Message

What I'm looking for

I want to continuously learn in AI/ML, turning research into real impact—building accurate voice and vision models, improving reliability with better algorithms, and collaborating on web products that solve everyday problems.

I’m a self learner and enthusiastic B Tech C.E student with skills in Python Programming, Web Development, Gen AI, and research in AI and Machine learning. I’m also familiar with Prompt Engineering Tools and I enjoy public speaking, constantly aiming to learn and grow in fast-moving technology.

My work includes an AI model for Voice DeepFake detection, achieving 92.8% accuracy using Librosa and MFCCs with AI and deep learning algorithms. I’ve also built a drone and swarm movement detection system using YOLO, and developed web projects using HTML, CSS, JavaScript, and MERN Stack—plus earned a winner 1st rank in a local business directory hackathon.

Experience

Work history, roles, and key accomplishments

AI Deepfake Voice Intern

Indian Army

Built an AI voice deepfake detection model to classify real vs. AI-generated audio, achieving 92.8% accuracy. Developed additional AI/ML systems for drone/swarm movement detection (YOLO) and AI-based Chinese-to-English document translation.

Speech Processing Model Evaluation Python Machine Learning Deep Learning Librosa Audio Signal Processing YOLO Computer Vision

Education

Degrees, certifications, and relevant coursework

Marwadi University

Bachelor of Technology (Computer Engineering), Computer Engineering

2022 - 2026

Activities and societies: Skills include Python programming, web development, genAI, prompt engineering tools; research and public speaking.

Pursuing a B.Tech in Computer Engineering at Marwadi University (Rajkot). Worked on Python, web development, and AI/ML projects, including ongoing voice deepfake research with reported 92.8% accuracy.