Mohammad Akram Hossain google scholar link

San Jose, CA. linkedin.com/in/mohammad-akram-hossain

Email: advanv@r.postjobfree.com github.com/shamol84/ Phone: 1-330-***-****

Summary

A professional machine learning engineer and data scientist specializing in time-series analysis, MLOPs, forecasting, optimization, data mining, deep learning, NLP, storage AI, anomaly detection, big data technologies, and Hadoop with 10+ years of experience. Technical Skills

● Languages/Version Control Systems: Python, R, Linux, shell script, Jira, GitHub, Bitbucket.

● Databases/Big Data Framework: SQL, NoSQL, Hadoop, Spark, AWS, GCP, Kubernetes, Docker, Sagemaker, Kubeflow, S3, Athena, Glue, BigQuery, Redshift, Beanstalk

● Deep Learning Framework: TensorFlow, Theano, Keras, PyTorch Work Experience

Senior Data Scientist/Machine Learning Engineer, Intel Inc. Santa Clara, CA Sep 2021- Present

● Built and implemented a machine learning tool to optimize data movement between DRAM and CXL. The machine learning algorithm has been integrated with Memtierd tool and Intel Analytics Accelerator (IAA). The implementation of this tool in Intel CPU improved cloud workload performance by 4~5%.

● Created and implemented an AI-based storage optimization algorithm to improve storage latency and reduce cost.

● Lead and implemented a prediction model to identify unknown workloads and noisy neighbors in cloud servers and recommend optimizations based on similar workloads.

● Implemented an optimizing algorithm to tune various workloads running in cloud servers which improved the performances of the cloud workloads by 5~8%.

● Developed a reinforcement learning-based leak detection method for the liquid cooling of the servers which can identify various levels of cooling leaks and provide suggestive actions to reduce server downtime.

● Tools used: Python, Kubernetes, Docker, AWS, Sagemaker, Memtierd Senior Data Scientist, Hanwha Q CELLS America Inc. Santa Clara, CA Apr 2019-Sep2021

● Developed an image processing tool to identify and quantify PV defects and cell cracks in real time during manufacturing.

● Created a real-time energy consumption and energy storage optimization model.

● Designed and implemented a big data mining and analytics infrastructure for energy and battery management systems.

● Lead a team to develop a customer-facing data monitoring platform and mobile App.

● Built an automated claim processing, response, and resolution system to automate the claim management process.

● Tools used: Python, PySpark, Hadoop, AWS, Flask, Django. Data Scientist/Machine Learning Engineer, Cisco Systems, Inc. San Jose, CA Aug 2017-Apr2019

● Lead and implemented an automated email response system, iReply, using natural language generation (NLG) through RNN deep learning algorithms. iReply helped reduce customer response time in less than 2 hours.

● Built and implemented a sales opportunity scoring model and customer journey prediction model for digital sellers.

● Developed product recommendation model for digital sales and marketing team.

● Tools used: Python, PySpark, Hadoop, Google Cloud, TensorFlow, Flask. Data Science Internship, Goodyear Tire and Rubber Company Cleveland, OH May 2017-Aug2017

● Developed a tool to collect real-time vehicle location data, video recording/image data, and vehicle onboard diagnostic data.

● Applied data science and analytics to derive insights from data, classify driving behavior, and provide real-time feedback. Machine Learning/Data Science Research Assistant, SDLE Research Center Cleveland, OH Aug 2012- May 2017

● Funds/Collaborators: Advanced Research Projects Agency-Energy; Cisco, Johnson Controls International (JCI)

● Developed and implemented a cloud-based statistical software for virtual energy audit tool known as EDIFES to reduce 20-30% building energy consumption. EDIFES is currently a startup at Cleveland.

● Developed unsupervised non-intrusive load disaggregation algorithms using the Hidden Markov model to identify equipment level energy consumption, behavior, and anomalies.

● Built and implemented an image processing tool to identify good PV and defective PV cells with 97% accuracy using machine learning techniques. The tool can identify PV cell cracks and dead PV cell areas and quantify their growth.

● Lead and built data acquisition, monitoring, and ingestion frameworks for power plants, weather, and building utility data.

● Develop a predictive model for real-world thermal behavior of microinverter with 92% accuracy from features extracted from weather and power sensors by applying sensor fusion and machine learning algorithms. Education

● Ph.D. in Mechanical Engineering, Case Western Reserve University, OH.

● MS in Mechanical Engineering, Case Western Reserve University, OH.

● BS in Mechanical Engineering, Bangladesh University of Engineering and Technology, Bangladesh.

Contact this candidate