𝘿ΞVΛNIK
class Devanik:
def __init__(self):
self.name = "Devanik"
self.education = "Electronics & Communication Engineering @ NIT Agartala '26"
self.interests = ["Reinforcement Learning", "Generative AI", "Deep Learning"]
self.current_focus = "Building AI systems that solve real-world problems"
self.achievement = "1st Place in Space Exploration Hackathon by ISRO STIC"
self.applications_built = 25+ # Using Streamlit with robust ML models
def daily_routine(self):
return ["Learn something new about AI", "Solve DSA problems", "Work on ML projects", "Contribute to open source"]
def life_philosophy(self):
return "AI, when harnessed correctly, can increase human lifespan to infinity""The future belongs to those who believe in the beauty of their dreams." - Eleanor Roosevelt
 Dec 2024 - Present · 5 mos |
• Working on "A Stochastic Approach to Reconstruct Gamma-Ray Burst Light Curves" under Prof. Maria Giovanna Dainotti • Applying models like Willingale Model (W07), Broken Power Law (BPL), and Gaussian Processes • Analyzing GRB light curve data to better estimate key parameters such as plateau end time, flux, and decay indices • Exploring ways to reduce uncertainties in GRB measurements for cosmology and theoretical studies • Gaining experience in statistical and machine learning techniques for astrophysical research |
 Nov 2024 - Feb 2025 · 4 mos |
• Developing Planto.ai Copilot, an intelligent assistant designed to enhance user productivity • Data Preparation: Cleaning and organizing datasets for training, validation, and testing • Model Building: Implementing neural network architectures tailored to the copilot's objectives • Training and Validation: Using advanced algorithms and monitoring performance • Fine-tuning: Optimizing models and hyperparameters to enhance predictions and prevent overfitting • Integration and Testing: Collaborating to integrate trained models into the platform |
 May 2024 |
• Completed sustainability consulting simulation • Developed a comprehensive fitment matrix for client needs • Created data-driven sustainability recommendations |
 May 2024 |
• Analyzed complex datasets to uncover business insights • Applied statistical modeling techniques to identify patterns • Created compelling data visualizations for stakeholder presentations |
- Cancer Cell Classification: Implemented and compared multiple classification algorithms
- Achieved 96.8% accuracy with ensemble methods
- Optimized hyperparameters using grid search and cross-validation
- Applied feature engineering to improve model performance
- Customer Segmentation: Applied K-means clustering to identify distinct customer groups
- Discovered 5 key customer segments with unique purchasing behaviors
- Used silhouette score to determine optimal cluster count
- Created actionable insights for marketing strategies
-
Stock Price Prediction: Built RNN models to forecast market trends
- Implemented LSTM and GRU architectures with time-series data
- Achieved 32% lower RMSE than traditional forecasting methods
-
Image Classification: Developed CNN models for computer vision tasks
- Fine-tuned advanced architectures like VGG16 and ResNetV2
- Utilized transfer learning to maximize performance with limited data
|
|
|
- IBM Data Science Professional - Mastering data analysis and machine learning
- Google Machine Learning - Advanced ML techniques and TensorFlow implementation
- Microsoft AI Fundamentals - Essential concepts in artificial intelligence
- TCS Machine Learning - Enterprise-grade ML solutions development
- FreeCodeCamp Data Analysis - Comprehensive data analysis techniques
"I believe that AI, when harnessed correctly, can increase human lifespan to infinity. The intersection of machine learning and healthcare is where I aim to make my greatest contribution."