Department of Computer Science & Engineering (CSE)

Linkedin: https://www.linkedin.com/in/icecoldmartini/

GitHub: https://github.com/IceColdMartini

Room No: 323, 3^rd Floor, Faculty of Electrical & Computer Engineering (FECE) Tower – 3, MIST.

Mirpur Cantonment, Dhaka-1216, Bangladesh.

EDUCATION

Khulna University of Engineering & Technology            Jan 2020 – Sept 2025

BSc in Computer Science and Engineering, CGPA: 3.86/4.00           Khulna, Bangladesh

• Advisor: Professor K. M. Azharul Hasan
• Senior-year Thesis Supervisor: Professor Sk. Imran Hossain

Notre Dame College                                                                                       May 2017 – July 2019

Higher Secondary Certificate, GPA: 5.00/5.00     Dhaka, Bangladesh

Ideal School & College                                                                                 Jan 2012 – April 2017

Secondary School Certificate, GPA: 5.00/5.00     Dhaka, Bangladesh

TEACHING EXPERIENCE

Lecturer | Military Institute of Science & Technology      Nov 2025 – Present

Department of Computer Science and Engineering      Dhaka, Bangladesh

• Courses Taken:

• CSE 106: Structured Programming Language Sessional 
• CSE 316: Digital System Design Sessional 
• CSE 364: Software Development Project - I 
• CSE 444: Pattern Recognition Sessional 
• EECE 280: Digital Electronics and Pulse Technique Sessional 

INDUSTRY EXPERIENCE

Software Engineer | Intelsense.AI           July 2025 – Present

Dhaka, Bangladesh

• Key Responsibilities

• End-to-End ASR, TTS Development, Fine-tuning, Scaling
• Handling fast inference and load balancing for local hosting
• Agentic AI, RAG Pipelines, Agent Tool Dev & Deployment
• Product Development
• UnisenseAI, a one-stop solution for an unified response system paired up with complete lead and transaction automation with large-scale recommendation models.
• SenseForm, a central privacy-protected form automation system for Bangladesh’s national banking for home banking and navigation with multilingual support.
• AirVoice, an ASR-TTS coupled system developed for improving Huawei’s Customer Service Op.

RESEARCH INTEREST

• Domain: CoreML / Statistical ML, Reinforcement Learning, LLM Alignment, LLM Reasoning, Model Compression, Knowledge Distillation
• Research Questions of Current Interest:
  • Can we achieve pure multimodality for Small Language Models (SLMs)? If yes, then how?
  • Can we articulate a more optimized flow to train context-rich student Small-VLMs from proprietary LLMs / VLMs?
  • End-device deployable privacy-preserving SLM development in ClinicalAI setting to enable edge-computing

AWARDS & GRANTS

• Dean’s List Award | KUET Faculty of Electrical and Electronics Engineering

• Recipient of Fresher, Sophomore, Junior and Senior awards for the sessions 2019-2023
• Academic Excellence Technical Grant | KUET Department of Computer Science and Engineering
• Recipient for Sophomore, Junior, and Senior years for Sessions 2020-23

TECHNICAL SKILLS

• Model Compression & Optimization
  • Quantization [INT8/FP16], Pruning [Structured/Unstructured], Knowledge Distillation, Low-RanK Factorization, Model Profiling & Benchmarking
• Advanced Python
  • PyTorch, TensorFlow, Hugging Face Transformers, ONNX, Scikit-learn, NumPy, Pandas
• Multimodal AI
  • Vision-Language Models, Cross-Modal Fusion, Multi-Modal Transformers, Attention Mechanisms
• MLOps & Deployment
  • Docker, K8s, Model Versioning, CI/CD Pipelines, Performance Monitoring, A/B Testing
• Model Alignment & Fine-Tuning
  • SFT [Supervised Fine-Tuning], DPO [Direct Preference Optimization], Knowledge Distillation for Alignment, Safety-Preserving Compression
• Healthcare AI Development
  • Medical Data Pre-processing, HIPAA-Compliant Systems, Clinical Dataset Handling, Privacy-Preserving ML