Extracting mechanical properties for micro pillars from video data• Utilized deep learning techniques to extract features from micro pillar compression videos• Extracted physics-based features like velocity and angle using computer vision techniques (sparse optical flow)• Predicted force displacement curve from the features using LSTM based networks to achieve Intersection over Union (IOU) of 0.91• Won Best Research Award at Carnegie Mellon Symposium 2021Reverse engineering Computer Aided Design (CAD) models using deep learning• Successfully reverse engineered CAD models using Transformers and Long Short Term Memory (LSTM) Networks• Implemented Point-Net architecture as a feature extractor from 3D point cloud of CAD parts for sequential operation prediction• Generated a dataset of 550 CAD parts with multiple features and their corresponding labels to train the deep learning modelsMetal 3D printing Spatter Detection using Deep Learning and Image Processing• Designed custom U-net architecture to perform pixelwise spatter classification and obtained 97.3% accuracy on the test set• Formulated custom image processing technique based on pixel brightness (PBT) to avoid expensive manual pixel data labeling• Improved model generalization using data augmentation to perform better on unseen component images• Utilized model compression techniques to reduce inference time by 30.8% and model size by 41.3%

• Mentoring students with semester-long projects on Computer Vision and Natural Language Processing• Designing assignments and midterm exams • Responding to questions asked by students on discussion forums such as Piazza• Holding weekly office hours

• Created a generalized deep reinforcement learning template to apply it to design optimization and material modeling problems• Implemented sequence-to-sequence models to predict material properties of different non-linear materials• Trained Radial Basis Function (RBF) networks for multi-constraint design optimization• Reviewed literature, summarized progress and best approaches on application of deep learning to mechanical design optimization

• Collaborated with all team leads to ensure effective functioning of the organization• Guided all sub teams to make better engineering decisions• Aided in designing and fabricating all aspects of the race car• Acquired 12th position overall in FSAE Michigan 2019 Competition out of 120 teams globally

• Provided advanced non-linear finite element simulation support to all tire design groups in North America • Worked on developing new and more advanced simulation tools for predictive technologies group• Aided in development of Cooper’s in-house Computer Aided Development Software• Developed training modules for Tread Stiffness simulations• Established well-documented methods for adding detailed tread geometry to simulation• Developed Python Scripts to automate better simulation data representation

• Led and contributed to a team selected by SpaceX to design and manufacture the future mode of transport• Aided in designing, manufacturing, and testing all components of the pod• Finished among top 22 teams out of 1200 teams (2016) and top 47 out of 1000 teams (2017)

• Programmed an algorithm in MATLAB to optimize build orientation to manufacture a part via any additive manufacturing process• Implemented an algorithm that considers several criteria, such as accessibility and minimization of support structure volume

• Completed a number of Abaqus courses involving Abaqus/Standard, Abaqus/Explicit, Heat Transfer, Coupled Eulerian-Lagrangian Approach and Contact Convergence• Successfully completed real life simulation problems including Bottle top load, Drop analysis and bottle squeeze analysis• Worked on modeling simulations involving fluid-structure interaction• Optimization of design of flexible packaging products (Cardboard)• Worked on Heat transfer analysis and drop problems for hand held consumer cameras• CAD modeling of various products

• Created and verified mechanical layouts, safety layouts, air layouts and detail drawings on various projects • Planned equipment layout, material flow, and measures to maintain efficient and safe utilization of resources and equipment on Amazon projects • Collaborated with other engineers to clarify and resolve problems found during execution; prepared design modifications, as appropriate• Worked with other engineering disciplines in the development and application of material handling equipment and solutions as required in support of project needs. • Integrated findings and concepts into design recommendations on projects assigned