MRIDUL BANIK

Fort Collins, CO, ***** +1-781-***-**** advm9f@r.postjobfree.com

Education

Work Experience

Machine Learning Engineer May 2022– Present

Leela AI, Newton, MA, USA

• Developing 10x faster inference on Custom Dataset by only sacrificing 2% accuracy on object detection using finetuned yolov5 and yolov7

• Developing module to invoke retraining from frames of undetected objects Machine Learning Engineer January 2020– July 2021

Apurba Technologies Ltd, Dhaka, Bangladesh

• Lead two Teams in Bengali OCR Project

• Deploying, Managing, validating Models and Applications in AWS Server (Ec2 Instance) with Docker

• Developing CI/CD Pipeline, Restful API, building and managing datasets, and performing exploratory data analysis

• Writing Reports (i.e. SRS Report, Inception Report) as deliverables Software Engineer August 2018 – December 2019

BRAC IED, Dhaka, Bangladesh

• Developed and maintained E-Learning platform using PHP, Google Cloud for 500 users

• Made traditional paperwork 50% faster by automation with JavaScript, Python, open-source application Graduate Teaching Assistant – [Software Engineering, Discrete Math] August 2021 – May 2022 Colorado State University, CO, USA

• Tutoring 80+ students per week

• Evaluating programming assignments, quizzes and assisting in one-to-one consultation Projects & Publications

• Analyzing Relationships among SVI index, Crime Cases and Covid Cases using Apache Spark and Scala [github repo]

• Accomplished 97% accuracy in URL - base phishing detection with 7 transformer networks [github repo]

• Achieved 92% accuracy and 99% recall score on detecting Pneumonia from chest X-ray images by transfer learning and custom CNN model [github repo]

• Achieved 97% accuracy in Bengali word segmentation by Differentiable Binarization model on custom built dataset

• Created a dashboard with Tableau that identify and prevent fraudulent transactions using data visualization and machine learning techniques to identify patterns and anomalies in transaction data..

• Implementation of FDNN model (Random Forest+Feed Forward Net) on Influenza Patient dataset with 94% accuracy detecting Influenza. Published in IEEE ICTAI & IEEE Conf. on Big Data

• Built Synthetic (5M Images) and Gold standard (8M images) corpuses for Bengali OCR with heuristic segmentation. Published Two papers in Springer [1, 2]

• Developed Bengali speaking robot using Google Translate API, Text to Speech API, Python ChatterBot Library Servo Motor, LED Light, Raspberry Pi. Published in IEEE Conf. (ICIEV) and (icIVPR)

• Developed Class Attendance System from weight & height using Raspberry Pi, Weight Sensor, RFID, Sonar Sensor, LED, Firebase Cloud, Python. Published in IEEE Conf. (ICIEV) and (icIVPR) Skills

- Java, Scala, Python, Shell Scripting - AWS, GCP, Sagemaker - Docker, Kubernetes

- Flask, Fast API, PHP, Postman - DVC, AirFlow, MLFlow, KubeFlow - Keras, Tensorflow, Pytorch

- MySQL, SQLite, Firebase

- Tableau, Scikit-learn, Pandas, Matplotlib - OpenCV, PIL, HuggingFace MS in Computer Science August 2021– May 2023

Colorado State University, CO, USA

CGPA - 3.39/4.0 (Studying)

[Courses: DBMS, ML, Distributed System, Big-Data, NLP, Security, Image Computation, Fault-Tolerant Computing] BS in Computer Science and Engineering May 2014– April 2018 BRAC University, Dhaka, Bangladesh

CGPA – 3.71/4.0 Mentions in VC’s List and Dean’s List Google Scholar

linkedin.com/in/mridul-banik/

github.com/mridulBanik112