Manmohan S.

Experience

• 

  Jet Airways

  May 2013 - Jul 2013

  Intern

  • Exposed to the engines, structures and mechanical systems of the Boeing 737 aircraft• Involved in technical services, planning and quality control department

• 

  Indian Institute of Technology, Kanpur

  Apr 2015 - Jun 2015

  Intern

  • Calculated center of gravity of aircraft for various number of passengers and sensor calibration of aircraft control surfaces for the Cessna Skylane aircraft• Evaluated static aircraft performance parameters including zero-lift drag co-efficient, ostwald efficiency factor, induced drag co-efficient for the Piper Saragota aircraft

• 

  University of Cincinnati College of Medicine

  Jan 2019 - Aug 2021

  Information Technology Analyst

  • Pitched and developed the use of NLP to reduce the response time for employees for critical requests by categorizing the requests into two different categories being ‘trivial’ or ‘nontrivial’• Utilized the Bidirectional Encoder Representation for Transformers (BERT) to train the model faster and achieved an accuracy of 98%• Debugged software issues during pandemic with Virtual Examinations for Medical students including NBME certifications and Honorlock • Maintaining software and hardware of the computers and multiple laptops in library as well as verifying the constant updates pushed through the server• Assisting medical students with their software, VPN and computer hardware issues at any given point of time (at risk) Show less

• 

  University of Cincinnati - College of Engineering and Applied Science

  Aug 2019 - May 2021

  Machine Learning and Deep Neural Network Experience:Center for Global Design and Manufacturing (Additive Manufacturing) Laboratory, UC Applied Transfer Learning (Fine tuning) to predict Sign Language images (CNN) • Applied MobileNet model on unseen data, in-turn accuracy attained - 91.67 % on the validation set • Developed a Transfer Learning model to predict different set of images using the VGG16 modelBuilt a Convolutional Neural Network (CNN) for multi-image classification on Intel dataset • Preprocessed the images using VGG16 model to improve the prediction accuracy • Plotted confusion matrix, performed manual error analysis and tuned the hyper-parameters to evaluate the performance of the classification model and improve its accuracyAnalyzed weekly corn price variation using data between 2015-2017 (LSTM) • Built a Long short-term memory (LSTM) neural network to predict the future variations of Corn price • Algorithm created can be modified for use in any kind of time series data prediction Show less

  • Graduate Research Assistant

    Aug 2017 - May 2021

  • Graduate Student Researcher

    Aug 2019 - Apr 2021
• 

  CattleQuants

  May 2021 - Aug 2021

  Machine Learning Engineer

  • Conducted image classification of cattle data, segregating them into different breeds, using the InceptionV3 model on imagery obtained from drones. Achieved exceptional precision and recall rates of 96% and 97% respectively across all classes.• Employed techniques such as Up sampling and Down sampling to address imbalances in the dataset, resulting in improved model performance and robustness.• Implemented Gradient-Weighted Class Activation Mapping (Grad-CAM) to generate visual explanations of model decisions, enhancing interpretability and understanding of classification outcomes. Show less

• 

  CattleQuants

  Oct 2021 - now

  Trained a GAN-based semantic segmentation model to accurately count cattle on ranches/feedlots.Developed a GAN-based semantic segmentation model utilizing Multiband input of RGB and Thermal imagery to count cattle.Implemented a GAN-based semantic segmentation model using Semi-Supervised learning on RGB ImageryAuthored a white paper on dissection of conjoined blobs using Machine Learning models.Mentored Machine Learning Engineer InternsConducted experiments using drone and created Orthomosaic and RGB point cloud on photogrammetry software such as Pix4D Mapper. Correlator, TerraSolid, ArcGIS, Global MapperSpearheaded the installation of drone in a box solution at Five Rivers feedlot and ran LiDAR experiments to estimate the weight of commodities at this location and fine-tuned the various parameters for better accuracy. Processed the LiDAR point cloud on LiDARMill software. Show less

  • Machine Learning Engineer II

    Dec 2023 - now

  • Machine Learning Engineer

    Oct 2021 - Dec 2023
• 

  Quantpoint

  May 2022 - Nov 2022

  Machine Learning Engineer