Work Portfolio Piece
By: Vikas Chinchansur (Vicky)
This page details the overall experience of working as an Intern at CSIRO. Later, I also describe the skills learnt and Achievements in the work. If you are looking for more details of the Anomaly Detection System Development Project, please click on the 'About Project' option in the menu, on the top left.
Skills
​Studying bachelors in Computer Science and having worked for four years as a full-stack developer, I must say, I possessed a skill set belonging to a service domain only. Pursuing Masters at research-based university like ANU, and now the industry experience of working as an intern at an institute like CSIRO, have helped me in expanding my skills on an exploration of new technologies in this field. Below is the list of the skill learnt in this process:
-
Communication skills: Improved Communication such as initiating talks, communicating effectively with members of the team, slowing down the speech, enunciating clearly.
-
Flexibility Skills: Getting self comfortable working remotely from home.
-
Presentation Skills: Scheduling, Leading, presenting and convening meetings with stakeholders and key members of other companies involved indirectly in the project.
-
Time Management: effectively scheduling and managing time for internship work in-corresponding to other works and tasks of life.
-
Research Skills: Learning of different research techniques and methods to carry in the project.
-
Creative Thinking: Strategics to think out of the box, to resolve the conflicts or approach the solution.
-
Listening Skills: This is an essential skill that I lacked. This project has helped me to improve on it continuously.
-
Positive Attitude: Having a positive attitude of can-do towards the tasks to be completed.
Experience
The Australian National University, as part of its Computer Science & Software Engineering Internship academic course (COMP8830), provides a breakthrough opportunity to its students looking for ICT industry experience. I was fortunate to have been selected as an intern by Australia's National Science Agency, CSIRO. Might be, having an active AWS Cloud Practitioner certification was a plus point. It was an amazing and completely different experience working here.
​
The knowledge transfer session on the project, The Anomaly Detection System Development, had kicked off a week before the start of the semester. Initially, it took a while to completely onboard us for the project, by granting us all the required access of the project websites, AWS accounts and Meeting platforms. It took more duration than expected, almost four weeks, to understand the basic concepts of the project. Later, I had to analyse the existing code of filters hosted on the cloud, to familiarise how different AWS services interact with each other. During mid-semester break, there was an additional task setting up the Anomaly Detecting Project on the local machine, by installing the open-source Anaconda Distribution. The main challenge was to resolve the package dependencies on various libraries such as Keras, TensorFlow, etc. With help from other interns in the project, Supriya Kamble, I was able to finally fix the issues after a span of a week.
​
The actual challenging work on the project started after the mid-semester break. There were many different methods and process to host the Machine Learning models onto the AWS cloud platform, and I was confused with the vast variety of options available. With the help and guidance of a new teammate, Jevy, having shown the directions, I was able to progress on the work to be carried out.
​
After having decided what methods are to be worked on, the challenge was to fix the issues faced and make the method under test work. I was stuck with the Method-1 (Call an Amazon SageMaker model endpoint using Amazon API Gateway and AWS Lambda) for a longer time, an issue related to AWS console, where I was working on fetching the output results from the AWS Sagemaker endpoint. Upon help from AWS Support, I was able to fix this issue. After getting familiar with the system issues working on method-1, it was easy to work on Method-2 (Creating a machine learning-powered REST API with Amazon API Gateway mapping templates and Amazon SageMaker).
​
In the last second week of the internship, I finalised the new design of the end-to-end architecture of our project to be hosted on AWS Cloud. Before, starting with the actual implementation work, we thought its better to discuss the proposal with the AWS Solution architect, to verify the design and get some inputs on improvements and feedback if any. We had a fruitful meeting, wherein we got some useful inputs on including AWS S3 Event Notification process, to reduce the wait-time of Lamda functions. As a result, this will also help us minimize the monthly billing of the project. In the final week, I had worked on writing the detailed report consolidating the results of all the research work and recommendations. Link to this report is specified in below Deliverables section.
Deliverables & Achievements
-
Deliverable-1: Designing the current architecture of the entire system.
-
Deliverable-2: Designing the new architecture of the system to include the new module, ML models onto the AWS cloud platform.
-
Deliverable-3: Prepared a report detailing all the results of trails of research methods tested. Along with conclusions and recommendations to the tasks to be carried out near-future on this project.
-
Statement of Work (SOW): As part of the internship course assignment, had an opportunity to work on SOW for the project. Understood the role of different stakeholders and their expectations. Learned to prepare project milestones and schedule, by working on the Gantt Chart. Real-time experience working on risk management and cost analyses of the project.
-
Appreciations: Received appreciation and applause from the AWS Solution Architect (Cain Hopwood, Amazon) and Project Manager (Peter Fitch, CSIRO) for the finalised architecture design for the prosed project.
-
Extensive Research Techniques: Carried out extensive search on the methods that are compatible with the project. Trailed and tested, three different ways of hosting the Machine Models on the AWS cloud platform, within a span of four weeks.
-
Presentation of Final Design: Delivered the presentation of final suggested Architecture design of the project. Please click on the link to view the PPT of the final presentation.
-
Internship Work Final Report: Here is the link of the 1622WQ Application Anomaly Detection System Development Research Report.
Get In Touch
Want to get in touch? We would like to hear from you. Fill the form below and we shall get in touch with lightening SPEED.
Mobile: +61-490488193
210, Northbourne Avenue,
Braddon, ACT - 2612