In our series Behind The Code, we reach out to the developers from the community to gain insights on how their journey started in the field of emerging technologies, the tools and skills they use and what they think is essential for their day-to-day operations.
For this week’s column, Analytics India Magazine caught up with Ashwin Murali, a Data and Infra Engineer at Memory.ai. Murali joined Memory.ai last month after working as an Associate Director of DevOps at Zoomcar.
At Memory.ai, Murali and his team create the Timely app which helps people automatically track time and see what they’ve been spending time on.
On the infrastructure side, Murali’s day-to-day work includes testing out the new infrastructure strategy as well as handling loads and auto-scale the application. As a data engineer, Murali works on TensorFlow and checks on optimising the codebase to get better throughput from the AI systems.
Murali came from a minimalistic mobile programming background which makes it difficult to understand Unix and Networking under the hood in his initial days as a developer. He overcame it by learning as well as getting his hands dirty with practical approaches. Also, during his early phase, the developer faced challenges in understanding the devices and hardware security in IoT which he eventually overcame by learning and implementing practical approaches.
Speaking about the cloud scene in India, Murali says that the acceptance of the cloud has been awesome with a lot of the front-runners playing at scale a lot of people could not think of a few years back. For example Freshworks, Razorpay, Flipkart, Zoomcar, etc. Murali says, “I’d definitely love to see an Indian equivalent of the AWS Gov Cloud – where every Indian Government entity is on the cloud, and it’s secure and audited and safe.”
On-Premise vs Cloud
According to Murali, the necessity and need to absolutely have an on-premise data centre are getting smaller now. On-premise has always been capital intensive for any organisation – where one has to pay a massive capital upfront for the compute resources. This is what the cloud took away by letting compute on demand and currently, the question is slowly shifting from On-Premise Vs Cloud to Vendor Vs Hybrid and Server Vs Serverless.
Common Challenges In Setting Up A Cloud Infrastructure
To this question, Murali mentioned three points or challenges which are commonly faced during setting up a cloud infrastructure.
- The first point he mentioned is the sizing. He described with an instance, “If the provider of choice is AWS, a lot of the “sane defaults” are covered. Security is on and is set to deny by default, setup can be automated to the dot. But no one tells you what size you need those systems to be. Too powerful, you’re burning cash, too less, you’re losing experience.”
- The second point is the use of terms and tools such as auto-scaling, serverless, microservices, among others without properly understanding the trade-offs that each term brings into the table.
- The third issue is the DevOps bandwidth which has been usually faced by organisations. Cloud needs a dedicated bandwidth to keep the systems running well and for this issue, there must be a dedicated DevOps person who will be checking out all these problems.
Murali’s Code, Toolkit & Skillsets
Being a DevOps and Infra engineer, Murali is a cloud-agnostic person and his favourite cloud platform is Amazon Web Service (AWS) while his favourite cloud management tool is happened to be Ansible. He said, “Nothing beats the simplicity of an SSH connection and a python script! With the advent of Hashicorp, Terraform, Packer, and Vault have become prominent members of the kit bag as well. But yes, my vote always goes to Ansible. A 1000 tools have come and gone, Ansible and Bash have always stood their fort.”
His Learning Resources
Murali said that his go-to resource is the ACM subscription which gives access to the entire O’Reilly Bookshelf. The O’Reilly bookshelf gives access to over 50000 books in the technology space which is incomparable. Then there is also the MIT courseware for any advanced topics. Currently, Murali is learning the MIT courseware’s Artificial Intelligence Podcast series by Patrick Winston and Mark Seifter.
Piece of Advice
Murali said, “ You should do what you’re good at and then when you see something in the tech space that’s new and interests you, you should take a deep look at it and start learning whatever you can about it. And when you do, you seek out opportunities in that space to use those newly acquired skills – could be talks, papers, or even a job. It’s an ongoing process that takes commitment and time.”
Provide your comments below
A Technical Journalist who loves writing about Machine Learning and Artificial Intelligence. A lover of music, writing and learning something out of the box. Contact: firstname.lastname@example.org