add learning day blog

_data/authors.yml

@@ -103,4 +103,10 @@ sanchit-ahuja:
   name: Sanchit Ahuja
   bio: ""
   github: sanchit-ahuja
-  website: https://sanchit-ahuja.github.io/
+  website: https://sanchit-ahuja.github.io/
+
+vinay-varma:
+  name: Vinay Varma
+  bio: ""
+  github: nvinayvarma189
+  website: https://nvinayvarma189.github.io/

_posts/2022-11-30-learning-day.md

@@ -0,0 +1,81 @@
+---
+title: The Learning Day Initiaitve
+date: 2022-11-30
+tags: [culture, learning-day, sticky]
+categories: [Machine Learning]
+image: assets/images/learning-day-blog/learning-day-dalle.png
+layout: post
+authors: [vinay-varma]
+latex: false
+---
+
+Greeting from the ML team at Skit!!!  Today we are announcing **The Learning Day Initiative**. 
+
+**Learning Day:** A dedicated working day where engineers are encouraged to spend time exploring their interests that may either help them level up in their jobs or help them advance in their careers.
+
+
+## Why is it important?
+The global landscape of AI & ML is ever-changing. With the exponential burst of research ideas and tools that the ML community produces every year, it is common for many ML practitioners to have a bulging heap of things and ideas waiting to be explored and experimented with. This often leaves ML practitioners overwhelmed.
+
+Especially at Skit, where we ship voice bots to our clients in record timelines, sometimes, ML Engineers are too busy solving the "urgent work". As we target to bring our go-live time further down, there is a risk of shrinking space for innovation and novelty. For a team that takes pride in shipping the best voice bots in the world, it is something that we cannot compromise on.
+
+Innovation doesn't always necessarily emerge from goal and deadline-bounded situations. Often groundbreaking products and tools sprout from playground experimentations. We look at the Learning Day initiative as a way of allowing our engineers to follow their curiosity and scratch their itch to build something that otherwise is less likely to happen with daily work.
+
+## How does it work?
+On every alternative Friday, we ask all our team members from our ML teams (ML Research, ML Product, and ML Platform) to announce their plans for the learning day.
+
+1. Folks are encouraged to get in touch with people from other teams to understand their workflows or form collaborations.
+2. Engineers spend their time reading/building things as per their plans.
+3. At the end of the day, participants are asked to produce a log of the things they learned.
+4. Feedback on what went right, what went wrong, and the scope for improvements to help the participants focus better is collected and acted upon.
+
+## What qualifies as learning?
+Learning can happen in any form. Some examples are:
+
+1. Re-implementing research papers.
+2. Trying to reduce the latency of our ML systems.
+3. Learning the tradeoffs of using different programming languages or frameworks for a specific task.
+4. Enrolling/Progressing in a course about niche ML concepts.
+5. Building tiny versions of popular algorithms and libraries.
+
+**Note:** It also does not necessarily be related to their job descriptions. Our engineers are free to explore seemingly unrelated things that help in advancing their careers. A couple of good examples can be:
+
+1. An Engineer trying to understand the principles of product management.
+2. An ML research Engineer learning about the System Architecture that powers our voice bots.
+
+_It can really be anything that an ML engineer and their manager have a consensus on._
+
+## What qualifies as productive output?
+We maintain a log to document and summarize the learnings of each participant for every learning day instance.
+
+Since the learning paths for each individual (and in general too) are very dynamic, we do not expect our engineers to always produce working prototypes at the end of the day.
+
+It can be a summary, a full-length blog, or a program that others can try out. Failure of experiments (which is also learning) is embraced too.
+
+## Examples of learning day logs
+While the format of the log can vary, we would like the participants to write key learnings. This is to ensure clarity and transfer of knowledge.
+
+**Good examples**
+1. Learned the difference between X and Y. I think X is better because of ABC reason.
+2. Built a tiny version of this python library X. For those interested, the code is [here](https://example.com).
+3. Discovered a library that can help write tests faster and cleaner. [Here](https://example.com) is a code comparison of a test with and without the library.
+
+## Sustaining the Initiative
+
+Driving forward such an initiative in a fast-paced environment comes with some challenges. We identified some major ones and have taken the following steps to ensure that the initiative is not side-tracked.
+
+1. **Bandwidth and Resource planning:** All business and product owners are informed and reminded about this initiative so that they can plan for tasks and assign work accordingly.
+2. **Daily work seeping in:** Everyone has the same day as the learning day. This avoids participants from submitting to internal/external pressure to complete regular tasks.  
+a. This is also why we are blocking an entire day as a learning day instead of taking out 20% every day.
+3. **High-priority escalations:** Not sure about Thanos but they are surely inevitable and they have the potential to turn the course around. In such cases, we believe that the participants should be aware of what is the priority. No hard rules for these. Participants should make the right decisions with help from their managers.
+
+## Road ahead
+We at Skit already have gone through a few iterations of the Learning Day. It's been quite exciting for us to see the richness and variety of learning paths of each individual.
+
+We desire to start posting a condensed form of our learning day logs henceforth. Why?
+
+1. It helps people outside Skit peek into the tech culture we cultivate.
+2. Sharing the learnings publicly may become the inspiration for someone to start tinkering with their projects.
+3. It can push the participants to do more. Don’t we all need some external motivation from time to time?
+
+Stay tuned to this blog series to track what our ML team is brewing!