📞 +1 (602) 862-8872 🌐 https://kunalkas.carrd.co/ 📧 [email protected]

• Arizona State University, AZ

  • Master of Science in Computer Science
  • GPA: 4.00/4.00
  • Duration: 8/2022 - 5/2024

• Vellore Institute of Technology, Chennai, India

  • Bachelor of Technology in Computer Science and Engineering
  • GPA: 8.73/10.00
  • Duration: 7/2015 - 4/2019

• Programming/Scripting Languages: Python, C++, SQL, R
• ML Framework and Tools: PyTorch, Pandas, Keras, OpenCV, Numpy, Tensorflow, Terraform, TorchGeo, CUDA
• Cloud Platforms and Other Skills: Azure, AWS, Docker, Qiskit, Ethereum, Blender, Unreal Engine, DVC, QGIS, HuggingFace, GEE
• Data Science Skills: Hypothesis testing, A/B Testing, Bayesian Learning, Statistical Programming, Spark, Airflow, MLFlow

• Duration: 8/2023 - 01/2024
• Currently working on detecting guns in CCTV streams and analyzing detections to trigger alarms. Developing robust neural inference engines using C++ and deploying them as containers on AWS GPU instances for efficient inference.

• Duration: 5/2023 - 8/2023
• Created and deployed a web crawler using Terraform IAC modules to commit results and search them using a GCS interface. Created the search bar, results page angular components and a NestJS backend to consume the search results from GCS.

• Duration: Jul 2019 - Jun 2022
• Worked on various projects including detecting bark-beetle-infested trees from drone footage/images for Sodra, creating a simulated forest environment using Blender and Unreal Engine, implementing time-series forecasting solutions, and developing solutions for obtaining Minutes of the Meeting using transformer-based language models.

• Created several optimized models including a custom vision model with high Receptive Field and low-parameter, a custom fast Resnet Architecture, a fully convolutional ViT, and RemoteDiffusion, an SD 1.5 model finetuned on RSCICD.

• Worked on projects such as detecting and segmenting cones on the Mars surface, mitigating racial bias amplification in the COCO dataset, and analyzing Reddit Pushshift db using PostgreSQL efficiently.

🚀 Looking for Collaborations? Feel free to connect with me on LinkedIn or follow me on GitHub.