Generative AI & LLMs
Building enterprise-ready language model workflows with agentic orchestration, retrieval, fine-tuning, and careful prompt systems.
LLM DevelopmentLangChainCrewAIPrompt EngineeringFine-tuning
জ্ঞান01
AI
with empathy
Machine LearningDeep LearningNatural Language ProcessingGenerative AIComputer VisionReinforcement LearningStatistical AnalysisMLOpsMachine LearningDeep LearningNatural Language ProcessingGenerative AIComputer VisionReinforcement LearningStatistical AnalysisMLOps
About
The story of
an ultralearner
I’m a passionate data scientist and AI researcher dedicated to leveraging cutting-edge technologies to solve complex problems. With a strong academic foundation in statistics and big data analytics, I focus on translating data into systems that are useful, trustworthy, and human-centered.
My academic journey has given me an intense appreciation for the world of facts and technology. I’m consistently prepared to confront new challenges across generative AI, language systems, computer vision, and production machine learning.
“In Bengali, my name means ‘one who knows everything.’ I treat that as a daily invitation to keep learning.”
50+
AI Models
1M+
Data Points
10K+
Learning Hours
Expertise
Areas of specialization
Deep Learning & NLP
Designing neural architectures for language and speech, from transformer models to transfer-learning systems for low-resource languages.
PyTorchHugging FaceTransformersASRAttention
ভাষা02
Machine Learning Engineering
Shipping end-to-end pipelines with robust feature engineering, model evaluation, explainability, and production-minded deployment.
Scikit-learnMLOpsCloudFeature EngineeringAnalytics
কর্ম03
Selected Work
Featured projects
01
Quantifying Myself
A personal data analysis project exploring self-discovery through quantitative analysis of daily life patterns, habits, routines, and emotions. Using statistical analysis and data science techniques to uncover meaningful trends and correlations in personal behavioral data.
Data ScienceStatistical AnalysisEDA
02
Empowering Health and Safety at Aurum Mining Corporation
A machine learning project developed for Trilytics '23 Analytics Case Study Competition by PGDBA, IIM Calcutta. Achieved 95% accuracy using Random Forest for mining incident classification. Our team advanced to the second round of the competition.
Machine LearningRandom ForestData Science
03
Solving Grid World Problems using MDP
Implementation of Markov Decision Process (MDP) algorithms to solve grid world navigation problems. This project explores fundamental concepts of reinforcement learning through practical application.
Reinforcement LearningDeep Reinforcement LearningPython
Recognition
Honors & awards
Nov 2025
Guild Awards
EY — Associated with EY
Guild Awards for the out of box thinking and building a research prototype for the industrial problem related to drug discovery.
Aug 2025
Client Extraordinaire Award
EY — Associated with EY
Recognized by EY with the Client Extraordinaire award for outstanding contributions to generative AI initiatives. This recognition highlights impactful collaboration, innovation, and delivery excellence in driving successful AI projects that created measurable value for clients.
Research
Cyclone Damage Detection Dataset for Disaster Impact Assessment
Bidit Sadhukhan, Soumen Halder, Debajyoti Maity, Subhamoy Bhaduri, Rahul Bhattacharyya, Sagnik Dutta, et al.
Cyclones are among the most devastating natural disasters impacting human society. One of the critical challenges following a cyclone is the assessment of building damage. Essential information such as the extent, severity, rate, and types of damage is required for rescue operations, humanitarian aid, and post-disaster reconstruction. Remote sensing techniques play a crucial role in acquiring such damage-related data due to their non-contact nature, cost-effectiveness, wide coverage, and rapid response capabilities. Across various disaster scenarios, including cyclones and armed conflicts, accurate and timely data on building damage and population displacement remain essential for effective relief efforts. In this research, a novel dataset has been prepared by preprocessing raw satellite images, which have been manually annotated by an expert team. A baseline result has also been produced for this dataset using widely adopted computer vision algorithms. Satellite imagery serves as a valuable source of information for assessing damage in disaster-affected regions. However, the vast amount of data that requires analysis makes manual evaluation impractical. The proposed dataset and baseline results aim to facilitate advancements in automated cyclone damage detection and geospatial analysis.
Disaster ManagementSatellite ImageryComputer VisionDatasetCyclone Damage
Read publication Collaborate
Let’s build intelligence with purpose.
Newsletter
Stay updated
Subscribe for thoughtful updates on AI, data science, research notes, and new work from the archive.