Software Engineer -Working on Computer Vision, Machine Learning, Object Detection, Vertex AI (custom training)

- Incorporated Maven and Jenkins for building and testing our Java projects. Maven helped us manage project dependencies and streamline the build process, while Jenkins automated the testing and deployment phases.- Additionally, I was responsible for maintaining the services hosted on AWS. This involved managing EC2 instances, setting up Auto Scaling groups to handle fluctuating traffic demands, and utilizing S3 for storing artifacts made in the build process.- I handled IAM account management, ensuring that team members had the appropriate permissions to access AWS resources securely. Finally, I implemented AWS CodeDeploy to automate the deployment of our application to the AWS infrastructure.

- Inteligen was building a self service Cloud governance SaaS solution. The idea was to discover cloud resources in customer account and then overlay the NIST/SOC2 compliance model on top of that and show their current compliance posture.- Using Duo cloud mapper (https://github.com/duo-labs/cloudmapper), an open source cloud discovery tool, I made modifications to parse through the meta data of each resource and build Network topology and Parent/Child/Peer relationships between resources. Cloudmapper provides some relationships and grouping but that was not sufficient.- Discovery and topology data was graphed using Cambridge Intelligence graph library. - Added support to check for security risk - unencrypted EBS volume, unsecured security group, orphan objects and so on.- Added mapping of SOC2 controls to cloud resources.- Using PyJQ filters implemented 40+ API to evaluate security risks like EBS/S3 encrypted or not.

- Built a FinOps tool to provide fine grain accuracy into the AWS Cloud consumption. Many resources are created dynamically for short duration like QA kitchen sink and smoke tests, scaling and performance tests, chaos monkey tests, spot instances, incomplete cleanup jobs. These were uncontrolled and used to result in unexpected big expense. - Tool UI front end was JavaScript node.js and backend was python-based AWS account crawler, using boto-library. It establishes secure access to AWS account using Assumed role and store data in MySQL every 30 min. - It generated what’s changed and sent report to SRE leads. The report contained what/who/when/cost added. It also reported orphan objects like EBS volume, PVC & EIP. - Front end provided AWS account onboarding, authentication flow, dashboard and reports.- This tool helped Alkira SRE's to optimize their process and bring cost control.