Anushka de
Intern @Indian Statistical Institute, Kolkata
Signup · Get unlimited contacts
WORK HISTORY
Intern @Indian Statistical Institute, Kolkata
Kolkata, IN
Conduct research and experiments in machine learning, deep learning and federated learning for privacy-preserving models.Design and implement model training pipelines using Python, PyTorch and TensorFlow; perform data cleaning, augmentation and evaluation.Prototype edge-device deployments using Raspberry Pi and Arduino Uno to evaluate on-device inference and data collection.Implement secure data transfer and privacy-preserving training workflows using federated learning frameworks and best practices.Collaborate with supervisors and team to document results; prepare reports and present findings to research group.
EDUCATION
Heritage Institute of Technology
Bachelor of Technology - BTech, Computer Engineering
ABOUT ANUSHKA DE
I’m a passionate B.Tech (CSE) student currently in my 3rd year, Heritage Institue of Technology, combining academic excellence with hands-on research and intern at the Indian Statistical Institute. My primary research interests are machine learning, deep learning and federated learning, particularly building privacy-preserving ML systems that run on edge devices.At ISI I’ve contributed to data preprocessing, model prototyping, experiments with federated setups and integrating models with Raspberry Pi/Arduino prototypes. I’m comfortable with Python, PyTorch, TensorFlow and end-to-end pipelines. I also bring strong cybersecurity awareness for safe deployment of ML models (ethical penetration testing, secure data handling, vulnerability assessment).ICSE (Calcutta Girls’ High School) — 97%; AISSCE (DPS Ruby Park) — 91.2%. I’m open to internships, research collaborations and projects in Cyber-Security/ML/IoT/secure AI/Blockchain and cryptography. Message me to discuss ideas or collaborations. Let\'s Connect!
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.