Experience

- Futurewei Technologies Inc.

• ML Algorithms Intern (multimodal learning)
  May 2024 - Aug 2024

• ML Algorithms Intern (computer vision)
  May 2023 - Aug 2023

• Santa Clara, CA, USA

 -  Worked on vision transformers and multimodal learning.  

 - Designed a novel attention mechanism for vision transformer by fusing sparse and regional attention (published at T4V@CVPR24) 

 - Experimented with multimodal alignment on 12-lead ECG signals as a homogeneous multimodal time-series data (published at TS4H@ICLR24) 

- Experimented with a modal alignment scheme, robust against missing modality cases 

 - Was awarded the Intern Gold Award

- Proposed 2 patents on ECG multimodal learning and AR SLAM systems

Samsung R&D Institute Bangladesh

• Senior Software Engineer
  Feb 2021 - Jul 2021

• Software Engineer
  Feb 2019 - Feb 2021
  Dhaka, Bangladesh

- Worked on 5 A1 patents, related to health sensor data in smartphone, camera and AI
- Awarded the Best Rookie award for the year of 2019
- Awarded the Best Creator (Silver) award for contribution in patents (2019)
- Awarded the Best Inventor award for contribution in patents (2020)

   - Worked on developing web-based internal software for Samsung

   - Technologies used : Java, Spring, Angular, PostgreSQL

Ensembl

Summer Intern

May 2018 - Aug 2018

Cambridge, United Kingdom

Organization : Ensembl, EMBL-EBI

Mentors : Fergal Martin, Kostas Billis

Project : Transcript Comparisons, Refining Orthologous Relations to Transcript Level

Project Details: Orthologous relationships between genes allow us to infer about their significance and functionalities. If we study the properties and functions of one gene, those findings can be translated to the genes orthologous to it. However in Eukaryotes a fairly small number of genes are responsible for synthesizing a vast repertoire of proteins, which is mainly attributed to Alternative Splicing. Thus for Eukaryotes merely gene level orthology relation is not enough, we need to establish a deeper understanding in the Transcript level. In this project we studied the vast amount of orthologous transcripts between human and mouse, and developed an algorithm to rank the transcripts according to similarity or relevance. Finally we also developed a gui application that will allow the biologists or bioinformaticians to apply our findings in their research.

Summary : https://summerofcode.withgoogle.com/projects/#4855606393962496

GUI Application : https://github.com/EnsemblGSOC/nabil-gsoc-2018

Technologies used : BioInformatics, Python, BioPython, Flask, C++, Perl, Ensembl api, MUSCLE, InterPro, Html, Css, Javascript, MaterializeCss, JQuery