Skip to content

M Varshini

AI/ML Engineer • Bangalore • m******************@gmail.com • +91*******084 • drivetube.ai/•••••

Professional Summary

AI/ML Engineer with 0 years of experience specializing in Computer Vision, Generative AI and Natural Language Processing. Practical experience building real-time object detection and gesture-control systems, developing NLP and GenAI prototypes using Hugging Face, OpenAI and Gemini APIs, and deploying lightweight ML demos on Streamlit. Seeking an entry-level role to apply hands-on skills in model development, prompt engineering, data pipelines and prototype deployment.

Technical Skills

Programming Languages: Python,Java,GoLang
Web Technologies: HTML,CSS
Frameworks and Libraries: TensorFlow,PyTorch,Keras,Scikit-learn,Flask,Streamlit,OpenCV,Pandas,NumPy,Matplotlib,Seaborn
Databases: SQL,MongoDB
Cloud and DevOps: AWS EC2,AWS S3,Streamlit Cloud
Testing: Jupyter Notebook,Google Colab,Selenium,Playwright,Image Annotation,Video Annotation,Bounding Box Annotation,Dataset Validation,AI Data Quality Assurance
Data and Analytics: Machine Learning,Deep Learning,Neural Networks,Object Detection,Transfer Learning,Generative AI,Plotly,PowerBI
Tools and Methodologies: OpenAI API,Gemini API,Anthropic,Claude.ai,Git,VS Code
Computer Vision: MediaPipe,YOLOv8,Real-time Object Detection,Image Processing
NLP & LLMs: Hugging Face Transformers,NLTK,Text Classification,Summarization,Sentiment Analysis,Prompt Engineering,LLM Workflows

Work Experience

CodeGnan IT Solutions Pvt. Ltd.
AI/ML Trainee
Dec 2025 – Apr 2026
Training and prototyping role at an IT training solutions firm; developed ML, NLP and computer-vision prototypes and completed structured coursework in GenAI and model workflows.
Tech Stack: Python, TensorFlow, PyTorch, Keras, Scikit-learn, Hugging Face, OpenAI API, Gemini API, MediaPipe, OpenCV, PyAutoGUI, Pandas, Matplotlib, Git, Jupyter Notebook, Google Colab
  • Completed an intensive curriculum in Python, machine learning, computer vision, NLP, deep learning frameworks and prompt engineering, applying learned techniques to multiple end-to-end prototypes using TensorFlow and PyTorch.
  • Built data-preparation pipelines and automated EDA reports for multiple datasets using Pandas and Matplotlib, improving dataset readiness and repeatability for downstream model training.
  • Implemented core ML algorithms from scratch including manual gradient descent for linear regression to validate theoretical understanding and improve debugging skills for iterative model development.
  • Developed NLP applications such as sentiment analysis, text classification and summarization using Scikit-learn and Hugging Face Transformers; evaluated models with standard metrics and iterative error analysis.
  • Created GenAI prototypes integrating Hugging Face, OpenAI and Gemini APIs and applied prompt engineering to improve response relevance and reduce hallucinations during iterative testing.
  • Designed and implemented a Gesture Control System prototype using MediaPipe, OpenCV and PyAutoGUI implementing seven custom gestures for scroll, cursor control, drag-and-drop, pause, resume and screenshot to enable touchless HCI.
CodeGnan IT Solutions Pvt. Ltd.
AI Trainee — GenAI
May 2025 – Jun 2025
Short-term GenAI-focused training at an IT training firm emphasizing LLM workflows, prompt engineering and API integrations for generative systems.
Tech Stack: Python, Gemini API, Prompt Engineering, Hugging Face, OpenAI API, Jupyter Notebook, Google Colab, Git
  • Completed targeted GenAI training covering large language models, prompt engineering methodologies and Gemini API integration workflows to prototype conversational and generative applications.
  • Designed and implemented an AI-powered Timetable Generator using the Gemini API, automating schedule creation based on user constraints and demonstrating constraint handling via advanced prompts.
  • Authored and iterated prompt templates and conditioning strategies to improve output consistency, reduce irrelevant outputs and handle user constraints across scheduling scenarios.
  • Integrated Gemini API responses into a prototype front-end flow and developed validation checks to ensure generated schedules met user-specified constraints and preferences.
  • Performed qualitative and basic quantitative evaluation of generative outputs; used user-feedback loops and prompt refinements to increase relevance and reduce rework during testing.
  • Documented GenAI design patterns, API usage patterns and prompt-engineering best practices to support reproducibility and handover for future prototype development.
Computer Networks

Projects

Gesture Controlled Human–Computer Interaction System
Tools Used: Python, OpenCV, MediaPipe, PyAutoGUI, Computer Vision, Real-time Processing
  • Developed a real-time gesture control system using MediaPipe for 21-hand landmark tracking and OpenCV for webcam capture to enable touchless human–computer interaction.
  • Mapped seven custom gestures to OS-level actions (cursor movement, drag-and-drop, scrolling, screenshot, pause/resume) via PyAutoGUI, enabling accessible touch-free controls.
  • Optimized the pipeline for CPU-only environments to maintain real-time responsiveness without GPU acceleration, improving accessibility and deployability on typical consumer hardware.
Real-time Object Detection System
Tools Used: YOLOv8, Streamlit, streamlit-webrtc, OpenCV, Model Deployment
  • Built a live YOLOv8 object detection web app and deployed it on Streamlit Cloud to deliver browser-accessible real-time detection across desktop and mobile devices.
  • Integrated streamlit-webrtc with STUN servers to securely stream camera feeds from user devices and implemented dynamic UI controls for model selection and confidence thresholds.
  • Managed code and deployment workflows using Git and streamlined updates to the hosted demo to maintain reproducible deployments and rapid iteration.
Crime Rate Pattern Analysis — ML Study
Tools Used: Pandas, Scikit-learn, Streamlit, Data Visualization, Classification
  • Preprocessed and explored crime datasets to extract features and address data quality issues for modeling using Pandas and visualization libraries.
  • Built and evaluated classification models with Scikit-learn, diagnosing class imbalance and applying corrective techniques to improve predictive performance.
  • Developed an interactive Streamlit dashboard to visualize crime pattern insights across categories and geographies for exploratory analysis.
AI-powered Timetable Generator (Hackathon Project)
Tools Used: Gemini API, Prompt Engineering, GenAI, Prototyping
  • Created an AI-powered timetable generator during NeuralTornado AI Hackathon using Gemini API and advanced prompt engineering to automate schedule creation from user constraints.
  • Implemented iterative prompt strategies and validation logic to handle conflicts and ensure generated timetables met user-defined requirements.

Education

P.B. Siddhartha College of Arts & Sciences, Vijayawada (Affiliated to Krishna University)
B.Sc Computer Science with Cognitive Systems • Vijayawada, AP • 2023 – 2026
Coursework: Machine Learning, Data Science, Cognitive Systems, Artificial Intelligence, Python Programming, Database Management, Statistics & Probability, Computer Networks
Sri Vidya Arts & Science College
Higher Secondary — MPC Stream • 2021 – 2023

Certifications

Data Analytics Job simulation — Deloitte
AI-Prompting for Everyone — DeepLearning.ai
Data Science Fundamentals — Scaler
Getting Started with Generative AI — IBM
Machine Learning Foundations — AWS
Data Science using Python & Data Analytics — Great Learning

Achievements

  • NeuralTornado AI Hackathon — Built AI-powered Timetable Generator: Developed a GenAI prototype using Gemini API and prompt engineering to automate schedule creation based on user constraints.
  • 2nd Prize — Cyber Fraud Awareness Poster Design Competition: Used GenAI prompting techniques with LLM tools (Gemini) to create an award-winning poster demonstrating creative application of generative models.

Powered by Drivetube · Create your own profile at drivetube.ai