| layout | title | permalink |
|---|---|---|
page |
CV/Résumé |
/about/ |
Software people, love to code, practice software engineering ethics & create soft-stuffs.
Technology domain interest: Software Development (Python,nodejs), Machine Learning, Natural language processing, DevOps.
- I also love to hear & study people from different expertise like Philosophy, Mathematics, Physics, Cognitive Computational Neuroscience and Psychology
2011-2015 : BSc, Computer Science and Engineering; Chittagong University of Engineering & Technology, Bangladesh.
Current: Senior Software Engineer @ TigerIT Bangladesh Ltd
- Developed production pipeline for Speech Recognition, Multilingual Machine Translation, Text to Speech System. Improved Bengali Phoneme Transcription Standard, Language Modeling/Integration.
- Developed Commchat Android app's different ML/NLP & phonetic typing features.
- Improved KinderMate Android Kids Video app's automatic question answering, topic modeling, turn text response generation, text/document similarity(LDA, tf-idf),sentiment analysis, exceptional question/slang filtering features.
- Developed audio preprocessing system that extracts only audio part from noisy wavs that extensively reduces human effort to prepare quality speech data from collected noisy speech data.
- Fine tuned/trained wav2vec2, BERT,XLSR ,WMT/NMT, Fairseq(LSTM,self-attention), kaldi models.
- Developed an app(Android, Flask) for collecting/editing speech & text data, audio keyword search, audio events detection system.
- Improved image tampering classification(using VGG etc.), deepfake detection.
- Lead a data collection team for machine learning projects. Web scrapping, ETL.
Previous: Senior Software Engineer @ Samsung Electronics
- Developed Bengali Speech Recognition System with customized Bengali phonemes that earned a funded project from Samsung HQ(Seoul,South Korea) & awarded icon of the month (July-August 2018) for outstanding contribution to the development of Bangla ASR.
- Code review,codebase improvement for Samsung SDK & Tizen SDK,tracking/reproducing bug reports from app developers/user queries,find out solutions. This helps Samsung & Tizen Developers a lot in the process of monetization using Samsung & Tizen SDKs. Exposed with Zendesk, Confluence.
- Lead a team of 5 Masters Students from BUET with professor Muhammad Abdullah Adnan, PhD(California San Diego,UCSD, USA) on the development of large scale Bengali ASR for Samsung Bixby AI assistant, from data collection to data cleaning, deep learning architecture research, development, training, tuning & optimization.
- Developed Intelligent Chat Agent (web) for Real Time UI/UX Expression based on Chat Context (NLU, LDA Topic Modeling, Corpora Based Feature Mapping).
- Published for Samsung: "Customizing Grapheme-to-Phoneme System for Non-Trivial Transcription Problems in Bangla Language" on North American Chapter of the Association for Computational Linguistics (NAACL-HLT 2019), At Minneapolis, USA.
- Scrum, Smart documentation, agile activity, guided Samsung interns.
- Developed Samsung Apps(using Android SDK,Galaxy SDK,Tizen Mobile SDK,Tizen Wearables SDK), debugging & testing Samsung Galaxy Watch Studio.
- Developed IoT Client-Server app. Worked on IoT stack of IoTivity in Linux, Tizen & ARTIK board.
- Modified ColorPhun Game (Android) with Deeplearning AI to play between human & AI.
- Wearable Smart Watch Face development (using Tizen Native APIs, Web APIs & Samsung Galaxy Watch Studio) for Gear S2,S3,Fit2 etc.
- Samsung Developer advocacy, prepared dev tutorials, provided platform(C,C++,Web) support to Samsung devs.