About
I am an energetic person who loves to explore new things in the CS field. I believe learning should be a never-ending process and I feel that I have mastered this with the experience I gained over the years of my education. My aim is to utilize the expertise gained during my education and experience to develop cutting-edge technologies which would help me & the organization grow.
Work Experience
➔ Implemented an automated Machine Learning-based eKYC system, reducing
manual processing time from hours to <2 mins, resulting in savings of $72,000 a
year. Enhanced efficiency and accuracy in KYC verification.
➔ Spearheaded the successful implementation of an eKYC solution, resulting in an
impressive 80% reduction in KYC verification workload, leading to a substantial
decrease (55%) in KYC agent headcount.
➔ Implemented a robust face verification integration with selfie matching to ensure a
seamless and highly secure user authentication process before unlocking a car.
➔ Designed and deployed an efficient Extract, Transform, Load (ETL) pipeline to
store CCTV video footage data, facilitating seamless retrieval and processing for
computer vision model training and testing.
➔ Led the design and implementation of a containerized application architecture
facilitating seamless intercommunication between microservices and optimizing
system performance and scalability
➔ Crafted and deployed a comprehensive cloud infrastructure in AWS for ”ILLIYEEN”
leveraging diverse services to ensure robust and scalable operations.
➔ Successfully automated the employee exit notification process via secure email to
rectify a potential data breach, ensuring enhanced data security and compliance.
➔ Proficiently oversee and manage large-scale Extract, Transform, Load (ETL)
pipelines, ensuring seamless operations and timely delivery of valuable insights to
stakeholders
➔ Engineered a SOTA real-time face recognition attendance system integrated with
integrated time management capabilities, optimizing workforce tracking and
streamlining attendance procedures.
➔ Effectively implemented the Floating Point (FP) Quantization mechanism (FP32 →
FP16) to achieve a remarkable 50% reduction in memory usage, significantly
enhancing model loading speed and optimizing computational efficiency
Education
Projects
A small companion of your daily dhikr while you work or surf internet. Recite Tasbih everytime you open a new tab.
An android app which classifies Bangla Sign Image to text.
Awards
The top 80 contestants from across the nation in the Robi Datathon 2019.