Hridam Biswas

Hridam Biswas

AI Researcher | ML Engineer

Building Intelligent Systems for Real-World Impact

About Me

I'm an AI and Machine Learning Researcher with a passion for building deep learning models that solve real-world problems. Currently pursuing B.Tech at KIIT, I've contributed to research at KIIT's Official Research Department and published IEEE papers on semantic segmentation and autonomous systems. My work spans clinical prediction systems, spatiotemporal analysis, and production ML pipelines—delivering measurable impact across healthcare, autonomous driving, and education. I've optimized biological prediction models at ISTA (18% accuracy improvement), engineered learning prediction systems at Dhamm.Ai, and analyzed 100M+ records at Microsoft to reduce supply chain waste. I believe in AI systems that are not only powerful but also interpretable, safe, and beneficial to society.

Featured Projects

Sepsis Prediction using Ensemble Deep Learning

Built a clinical prediction pipeline combining XGBoost, multilayer perceptron, and logistic regression models with SHAP-based interpretability for transparent medical decision-making.

XGBoost PyTorch SHAP Clinical AI

🎯 High-stakes healthcare application with interpretable predictions

PyramidNet Semantic Segmentation

Designed a multi-scale convolutional architecture for autonomous driving scenarios, demonstrating robust generalization on real-world driving data. Published in IEEE.

PyTorch Computer Vision Semantic Segmentation

📄 IEEE Published Research

View Publication →

RLE XOR Trajectory Mining

Developed a spatiotemporal analysis system for autonomous vehicles to improve obstacle prioritization and path planning in complex driving scenarios.

Deep Learning Trajectory Analysis Autonomous Systems

📄 IEEE Published Research

View Publication →

Biological Prediction System Optimization

Optimized biological prediction systems at ISTA through structured hyperparameter tuning and deep learning architecture design.

TensorFlow Hyperparameter Tuning Computational Biology

📈 18% accuracy improvement | 22% latency reduction

Jurify

End-to-end deployed AI system that simplifies complex legal jargon into clear, actionable summaries. Utilizes optimized NLP models for high accuracy and fast inference.

NLP PyTorch Deployment

🚀 End-to-End Deployed | Legal AI Accessibility

MapMyWay

Deployed trip planning and navigation platform that creates optimized itineraries and provides real-time direction insights for any location globally.

Full-Stack Geolocation API Deployment

🗺️ End-to-End Deployed | Smart Trip Planner

Experience

Machine Learning Researcher

ISTA, Austria (Remote) | May 2025 - Jul 2025

  • Optimized biological prediction systems through structured hyperparameter tuning, increasing model accuracy by 18%
  • Architected deep learning models for biological datasets, improving prediction stability across experimental settings
  • Reduced inference latency by 22% through optimization techniques

Machine Learning Engineer Intern

Dhamm.Ai, Noida, India (Remote) | Feb 2025 - Apr 2025

  • Engineered student learning prediction models using behavioral data for data-driven academic forecasting
  • Implemented scalable training pipelines and feature workflows, reducing model retraining time by 30%

Data Analyst Intern

Microsoft, Bengaluru, India (On-site) | May 2024 - Jul 2024

  • Analyzed over 100 million records using SQL, Python, and Azure Data Explorer for supply chain optimization
  • Applied statistical modeling and machine learning to reduce spare parts inventory waste by 10-15%
  • Automated ETL workflows, improving data availability and reducing manual reporting effort

Machine Learning Engineer Intern

Ardent Pvt. Ltd., Kolkata, India (On-site) | Oct 2023 - Dec 2023

  • Fine-tuned transformer-based NLP models for feedback analysis with improved classification accuracy
  • Constructed deep learning models for enrollment and dropout forecasting, improving accuracy by 20%