Super excited to (re)start my Google journey as a Staff Engineer and this time I will be working on Ads Forecasting :-)
Generate viral LinkedIn posts in your style for free.
Generate viral LinkedIn posts in your style for free.
Arpit Bhayani
These are the best posts from Arpit Bhayani.
17 viral posts with 39,899 likes, 1,883 comments, and 228 shares.
3 image posts, 0 carousel posts, 0 video posts, 14 text posts.
đ Go deeper on Arpit Bhayani's LinkedIn with the ContentIn Chrome extension đ
Best Posts by Arpit Bhayani on LinkedIn
I felt like an imposter for the first few months at Google.
Whenever we switch to a new company or join a new team, the first few months are always scary and it happens to the best of us. New codebase, new processes, new people, new rituals, everything seems overwhelming.
Things get worse, when we put pressure on ourselves to perform and make an impact from day one. Remember, the initial learning curve at any company, regardless of your background, is steep.
So, how do we overcome this? I did the following things to overcome. Again, these are from my personal experience and not an exhaustive list.
1. acknowledge that you will not be the smartest one in the room
2. accept that you know nothing and it will take time to build the context
3. ask questions, it helps you learn the codebase and processes faster
4. observe others and see how they operate and navigate
5. work for a few extra hours, if required, to ensure you ramp up faster
So, the next time you feel that imposter syndrome, remember: it's a normal part of the process. Lean into the discomfort, ask those questions, and you'll be surprised at how quickly you become an integral part of the team.
#CareerGrowth
Whenever we switch to a new company or join a new team, the first few months are always scary and it happens to the best of us. New codebase, new processes, new people, new rituals, everything seems overwhelming.
Things get worse, when we put pressure on ourselves to perform and make an impact from day one. Remember, the initial learning curve at any company, regardless of your background, is steep.
So, how do we overcome this? I did the following things to overcome. Again, these are from my personal experience and not an exhaustive list.
1. acknowledge that you will not be the smartest one in the room
2. accept that you know nothing and it will take time to build the context
3. ask questions, it helps you learn the codebase and processes faster
4. observe others and see how they operate and navigate
5. work for a few extra hours, if required, to ensure you ramp up faster
So, the next time you feel that imposter syndrome, remember: it's a normal part of the process. Lean into the discomfort, ask those questions, and you'll be surprised at how quickly you become an integral part of the team.
#CareerGrowth
You'll never have to prepare for interviews if you are learning every day âĄ\n\nWhen you build a genuine interest in the domain you are in and in the work you do, interviews trim down to mere healthy discussions. Plus, genuine enthusiasm is contagious and it radiates through your responses leaving a positive impression.\n\nWhen you aim to become a better engineer, you will see cracking interviews becoming a piece of cake, a mere by-product.
New Year Resolutions changed my life, here's my story and the framework I follow đâ
December 2019: I decided to start my newsletter and push 52 essays in 52 weeks. I published one deep engineering article every single Sunday; no matter what!
The newsletter helped me increase my breadth, and become a better engineer. I also built a habit of reading research papers. Today 30000 engineers read the newsletter and become better engineers.
Newsletter: https://lnkd.in/d4qSXEdp
December 2020: I decided to earn 3 lac through a side hustle. I put in 2 months of insane effort in preparing the course and started my first batch of the System Design Cohort in March 2021.
Today, I am conducting the 14th cohort, and the course is enrolled by more than 1000 engineers from 14 countries. The course also helped me make great connections with senior engineers and leaders.
Course: arpitbhayani.me/course
December 2021: I set a goal to elevate the engineering ecosystem and hence I started putting out deep tech engineering videos on YouTube. Today, the channel has ~26k subs and growing.
YouTube: youtube.com/c/ArpitBhayani
So, what's my framework?
Every year, I reserve the last week of December to set 4 achievable goals for the next year. Every quarter, of the next year, I pick one and chase it.
Setting goals helps me to not digress and focus on things that matter.
Does this even qualify to be a framework? I don't think so; used it because it sounded fancy and intellectual.
Everyone can set goals, but very few hold the determinism to execute them.
So, this year focus on executing your goals; Trust me, the efforts are all worth it.
#AsliEngineering
December 2019: I decided to start my newsletter and push 52 essays in 52 weeks. I published one deep engineering article every single Sunday; no matter what!
The newsletter helped me increase my breadth, and become a better engineer. I also built a habit of reading research papers. Today 30000 engineers read the newsletter and become better engineers.
Newsletter: https://lnkd.in/d4qSXEdp
December 2020: I decided to earn 3 lac through a side hustle. I put in 2 months of insane effort in preparing the course and started my first batch of the System Design Cohort in March 2021.
Today, I am conducting the 14th cohort, and the course is enrolled by more than 1000 engineers from 14 countries. The course also helped me make great connections with senior engineers and leaders.
Course: arpitbhayani.me/course
December 2021: I set a goal to elevate the engineering ecosystem and hence I started putting out deep tech engineering videos on YouTube. Today, the channel has ~26k subs and growing.
YouTube: youtube.com/c/ArpitBhayani
So, what's my framework?
Every year, I reserve the last week of December to set 4 achievable goals for the next year. Every quarter, of the next year, I pick one and chase it.
Setting goals helps me to not digress and focus on things that matter.
Does this even qualify to be a framework? I don't think so; used it because it sounded fancy and intellectual.
Everyone can set goals, but very few hold the determinism to execute them.
So, this year focus on executing your goals; Trust me, the efforts are all worth it.
#AsliEngineering
What aches me is how everyone is just focused on DSA and, sadly, only for the sake of interviews and jobs. This is toxic and needs to change.
Computer Science is much more than just DS and Algorithms; there are many great subjects and verticals to explore and master.
No true innovation comes out of India because we study for jobs, not our innate curiosity and love for the domain.
By just learning a certain vertical, you limit yourself and restrain any inspiration you may take from other domains and verticals to build something unique and interesting.
Engineers not talking beyond DSA is just sad.
Computer Science is much more than just DS and Algorithms; there are many great subjects and verticals to explore and master.
No true innovation comes out of India because we study for jobs, not our innate curiosity and love for the domain.
By just learning a certain vertical, you limit yourself and restrain any inspiration you may take from other domains and verticals to build something unique and interesting.
Engineers not talking beyond DSA is just sad.
If you really want to learn, you'll find a way. If not, you'll find an excuse.
- being too busy - âdon't have the timeâ
- blaming work - âjob keeps me occupiedâ
- waiting for the right time - âwill start when things settleâ
- blaming external factors - âdon't have the right resources or the classic - blaming the education systemâ
Learning isn't about time - it's about priorities. If it truly matters, you'll find a way. If not, you'll find an excuse.
- being too busy - âdon't have the timeâ
- blaming work - âjob keeps me occupiedâ
- waiting for the right time - âwill start when things settleâ
- blaming external factors - âdon't have the right resources or the classic - blaming the education systemâ
Learning isn't about time - it's about priorities. If it truly matters, you'll find a way. If not, you'll find an excuse.
The best engineers I have worked with had learning built into their routines.
Try to take 30 minutes every weekday to read and watch technical things. Forming a habit around this will keep you curious, prepared, and upskilled. Here is an exhaustive list of things you can learn from [in order of my preference]
1. White papers
2. Conference videos
3. Newsletters and Blogs
4. Books
5. Open source codebases
6. Courses
7. Podcasts
Now that you have this exhaustive list, please don't worry about finding the best resources. Instead, get started with what you have because what truly matters is how consistently you can keep learning.
Pro tip: Try prototyping the core of every concept. That will give you a much deeper understanding of the concept.
#AsliEngineering #CareerGrowth
Try to take 30 minutes every weekday to read and watch technical things. Forming a habit around this will keep you curious, prepared, and upskilled. Here is an exhaustive list of things you can learn from [in order of my preference]
1. White papers
2. Conference videos
3. Newsletters and Blogs
4. Books
5. Open source codebases
6. Courses
7. Podcasts
Now that you have this exhaustive list, please don't worry about finding the best resources. Instead, get started with what you have because what truly matters is how consistently you can keep learning.
Pro tip: Try prototyping the core of every concept. That will give you a much deeper understanding of the concept.
#AsliEngineering #CareerGrowth
Apart from your regular work, spend some time doing maths, it will help you in the long run.
This one skill will become indispensable in the coming years and will keep you afloat.
This one skill will become indispensable in the coming years and will keep you afloat.
the best and the worst part of being a CTO,
in the morning, I am a backend engineer
in the afternoon, fine-tuned LLM for storytelling
in the evening, I set up a multi-region read replica
and now, working on the front end to improve UX.
but yes, entrepreneurship is easy.
in the morning, I am a backend engineer
in the afternoon, fine-tuned LLM for storytelling
in the evening, I set up a multi-region read replica
and now, working on the front end to improve UX.
but yes, entrepreneurship is easy.
Programming is easy, but Java makes it frustratingly difficult.
Instead of business logic, Java makes us focus on abstraction & extensibility. A simple IF is required to be modelled as multi-level abstraction, because SOLID, whyy!
Verbosity does not necessarily improve readability.
matlab, kuch bhi!
Instead of business logic, Java makes us focus on abstraction & extensibility. A simple IF is required to be modelled as multi-level abstraction, because SOLID, whyy!
Verbosity does not necessarily improve readability.
matlab, kuch bhi!
DiceDB is now open source and we have changed our license to AGPL 3.0; we believe that great infrastructure shouldnât be proprietary but a shared foundation for the ecosystem to build upon.
We are building a high-performance caching backbone for modern architecture, and switching our license to AGPL 3.0 marks a significant milestone in our journey - one that reflects both our technical clarity and our commitment to the open-source ethos
Our roadmap includes multi-tenancy, horizontal scalability, infrastructure abstraction, reactivity, observability, persistence, fault tolerance, cluster management, and a robust control plane.
We are also enhancing query capabilities and data structure support to meet the demands of AI-first applications - an approach poised to become the standard in the coming decade.
At our core, weâre just a group of passionate engineers pursuing an ambitious vision; join us and help shape and build the caching backbone of modern architecture.
read the announcement - https://lnkd.in/gjV7x4GX
meet the team - dicedb.io/team
community - dicedb.io/community
star the repository - github.com/dicedb/dice
#DiceDB #OpenSource
We are building a high-performance caching backbone for modern architecture, and switching our license to AGPL 3.0 marks a significant milestone in our journey - one that reflects both our technical clarity and our commitment to the open-source ethos
Our roadmap includes multi-tenancy, horizontal scalability, infrastructure abstraction, reactivity, observability, persistence, fault tolerance, cluster management, and a robust control plane.
We are also enhancing query capabilities and data structure support to meet the demands of AI-first applications - an approach poised to become the standard in the coming decade.
At our core, weâre just a group of passionate engineers pursuing an ambitious vision; join us and help shape and build the caching backbone of modern architecture.
read the announcement - https://lnkd.in/gjV7x4GX
meet the team - dicedb.io/team
community - dicedb.io/community
star the repository - github.com/dicedb/dice
#DiceDB #OpenSource
This afternoon, I got curious about image formats and started exploring the difference between JPEG and PNG, here's something that you will find fascinating âĄ
JPEG is a lossy format which means it throws away some data to get a higher compression rate, while PNG does not.
PNG tries to balance between the sizes of the image and the exactness (quality) of the image using a compression technique called DEFLATE which uses a combination of LZ77 and Huffman.
LZ77 finds the repeating patterns in the image data, storing a reference to them instead of the entire data itself. For example, if the next 100 pixels are red in color, instead of storing color for them individually, it combines the information and stores it as ânext 100 pixels are redâ; thus saving space when colors are consecutively repeated.
Huffman Coding helps in assigning shorter codes to more frequent colors and longer codes to less frequent colors, saving significant space.
This indicates that PNG is a great format for images that contain fewer colors and patterns; but if the image is vibrant and contains lots of patterns the file size might be larger than the JPEG counterpart.
⥠I keep writing and sharing my practical experience and learnings every day, so if you resonate then follow along. I keep it no fluff.
youtube.com/c/ArpitBhayani
JPEG is a lossy format which means it throws away some data to get a higher compression rate, while PNG does not.
PNG tries to balance between the sizes of the image and the exactness (quality) of the image using a compression technique called DEFLATE which uses a combination of LZ77 and Huffman.
LZ77 finds the repeating patterns in the image data, storing a reference to them instead of the entire data itself. For example, if the next 100 pixels are red in color, instead of storing color for them individually, it combines the information and stores it as ânext 100 pixels are redâ; thus saving space when colors are consecutively repeated.
Huffman Coding helps in assigning shorter codes to more frequent colors and longer codes to less frequent colors, saving significant space.
This indicates that PNG is a great format for images that contain fewer colors and patterns; but if the image is vibrant and contains lots of patterns the file size might be larger than the JPEG counterpart.
⥠I keep writing and sharing my practical experience and learnings every day, so if you resonate then follow along. I keep it no fluff.
youtube.com/c/ArpitBhayani
Docker is not the optimal way to run serverless workloads; hence, AWS Lambda or even an online judge like CodeChef, uses Firecracker as its execution environment.
This week, Iâm reading about Firecracker which is Amazonâs lightweight VM built specifically for serverless applications like AWS Lambda and Fargate. The paper covers how Firecracker supports thousands of short-lived containers and functions on shared hardware without sacrificing security or efficiency.
I skimmed the paper once, and it is pretty interesting to understand the story behind how Amazon found the right trade-offs between traditional hypervisors and containers and balanced both really well. Three things I found amusing are
- how Firecracker achieves sub-150 ms boot times
- how each VM uses just 3 MB of RAM, maximizing density on a single server
- how it leverages Rust and some key patterns to enhance security by design
definitely a good weekend read, if you have time :)
btw, admissions for my System Design December cohort are open. If you are SDE-2, SDE-3, and above and looking to build a rock-solid intuition to design any and every system, check out
arpitbhayani.me/course
#AsliEngineering #SystemDesign
This week, Iâm reading about Firecracker which is Amazonâs lightweight VM built specifically for serverless applications like AWS Lambda and Fargate. The paper covers how Firecracker supports thousands of short-lived containers and functions on shared hardware without sacrificing security or efficiency.
I skimmed the paper once, and it is pretty interesting to understand the story behind how Amazon found the right trade-offs between traditional hypervisors and containers and balanced both really well. Three things I found amusing are
- how Firecracker achieves sub-150 ms boot times
- how each VM uses just 3 MB of RAM, maximizing density on a single server
- how it leverages Rust and some key patterns to enhance security by design
definitely a good weekend read, if you have time :)
btw, admissions for my System Design December cohort are open. If you are SDE-2, SDE-3, and above and looking to build a rock-solid intuition to design any and every system, check out
arpitbhayani.me/course
#AsliEngineering #SystemDesign
You need not be passionate about tech to get started in tech.
It is okay not to be passionate and in love with computer science when you start in tech, but after some years, you need to love what you are doing to survive.
Because, if you are not passionate about something it would become extremely difficult to keep yourself updated with new languages, frameworks, tools, and even domains coming every now and then.
In the initial stages of your career, people pay you for the hands you have, and eventually, they pay you for the wits and depth you bring to the table.
#CareerGrowth
It is okay not to be passionate and in love with computer science when you start in tech, but after some years, you need to love what you are doing to survive.
Because, if you are not passionate about something it would become extremely difficult to keep yourself updated with new languages, frameworks, tools, and even domains coming every now and then.
In the initial stages of your career, people pay you for the hands you have, and eventually, they pay you for the wits and depth you bring to the table.
#CareerGrowth
Knowing when to switch roles or companies significantly impacts your career growth and trajectory and I have a simple 3P formula that can help you find the right time to switch.
1. Paisa (money)
2. Power (core competency growth)
3. Position (ladder growth)
At any company you are working at or switching to, you should get at least two of the three Ps. If you are getting fewer than two, it is time to switch.
1. Paisa (Money)
Monetary compensation is often a primary motivator for job change. Consider a switch if your current role does not provide enough or if the increments do not keep pace with industry norms. If the other two P's outweigh your average salary, it might be worth staying at the current company.
2. Power (Core Competency)
Power in this context refers to your growth in core competency and how close you are to becoming a subject matter expert in the domain you operate in. Aim to become a really good engineer, and a good job will always present you with opportunities to become one.
Assess whether your current role challenges you, introduces you to new technologies, methodologies, or projects, and ultimately contributes to your professional depth and breadth. Again, if the other two P's outweigh the lack of core competency, it might be worth staying at the current company.
3. Position (Ladder Growth)
The third P, Position, involves your upward movement in the org ladder. Your official title matters and it dictates the roles and responsibilities you have handled. Hence, an important criterion to decide if it is the right time to switch or not.
Assess if your current job provides a clear and actionable path for promotion and increases in responsibility. Stagnation can often lead to your future employer doubting your abilities and will negatively impact your career growth.
Again, if the other two P's outweigh the lack of ladder growth, it might be worth staying at the current company. Most people remain an L5 at Google is an example of this.
I always kept evaluating my situation every 6 months and kept over-optimized for two of the three Ps. For example,
1. at Practo, I optimized for Power and Paisa
2. at Amazon, I optimized for Position and Paisa
3. at Unacademy, I optimized for Position and Power
4. at Google, I optimized for Power and Paisa
My entrepreneurial stint has been about optimizing for Position and Power with a hope for a high gain in the third P in coming years.
To me, this has been a pretty structured framework to guide my thinking process, ensuring that my career decisions are both strategic and beneficial in the long run. Hope it helps you as well.
⥠I keep writing and sharing my practical experience and learnings every day, so if you resonate then follow along. I keep it no fluff.
youtube.com/c/ArpitBhayani
#AsliEngineering #CareerGrowth
1. Paisa (money)
2. Power (core competency growth)
3. Position (ladder growth)
At any company you are working at or switching to, you should get at least two of the three Ps. If you are getting fewer than two, it is time to switch.
1. Paisa (Money)
Monetary compensation is often a primary motivator for job change. Consider a switch if your current role does not provide enough or if the increments do not keep pace with industry norms. If the other two P's outweigh your average salary, it might be worth staying at the current company.
2. Power (Core Competency)
Power in this context refers to your growth in core competency and how close you are to becoming a subject matter expert in the domain you operate in. Aim to become a really good engineer, and a good job will always present you with opportunities to become one.
Assess whether your current role challenges you, introduces you to new technologies, methodologies, or projects, and ultimately contributes to your professional depth and breadth. Again, if the other two P's outweigh the lack of core competency, it might be worth staying at the current company.
3. Position (Ladder Growth)
The third P, Position, involves your upward movement in the org ladder. Your official title matters and it dictates the roles and responsibilities you have handled. Hence, an important criterion to decide if it is the right time to switch or not.
Assess if your current job provides a clear and actionable path for promotion and increases in responsibility. Stagnation can often lead to your future employer doubting your abilities and will negatively impact your career growth.
Again, if the other two P's outweigh the lack of ladder growth, it might be worth staying at the current company. Most people remain an L5 at Google is an example of this.
I always kept evaluating my situation every 6 months and kept over-optimized for two of the three Ps. For example,
1. at Practo, I optimized for Power and Paisa
2. at Amazon, I optimized for Position and Paisa
3. at Unacademy, I optimized for Position and Power
4. at Google, I optimized for Power and Paisa
My entrepreneurial stint has been about optimizing for Position and Power with a hope for a high gain in the third P in coming years.
To me, this has been a pretty structured framework to guide my thinking process, ensuring that my career decisions are both strategic and beneficial in the long run. Hope it helps you as well.
⥠I keep writing and sharing my practical experience and learnings every day, so if you resonate then follow along. I keep it no fluff.
youtube.com/c/ArpitBhayani
#AsliEngineering #CareerGrowth
1.6m concurrent users âper secondâ.
Peak Tech Journalism đđđ
Kuch bhi likhne ka matlab đ§ abhi bolenge distance bhi nikalo.
Peak Tech Journalism đđđ
Kuch bhi likhne ka matlab đ§ abhi bolenge distance bhi nikalo.
Has a CTO ever killed his production on a live podcast? âĄ
Subhash actually killed his production while we were recording this podcast to demonstrate a live production failover and that too on his bare metal infrastructure đ„
Recently, DukaanÂź ditched the public cloud and started hosting their application on bare metal instances. This was a very counter-intuitive move and hence invited Subhash to the #AsliEngineering podcast to discuss it in-depth.
Throughout the podcast, Subhash gave live demonstrations of failovers, observability setups, infra scripts, trade-offs, and much more. Watch this podcast and learn from the legend himself.
watch it here - youtu.be/vFxQyZX84Ro
Do spread the word and let us know what intrigued you the most.
Subhash actually killed his production while we were recording this podcast to demonstrate a live production failover and that too on his bare metal infrastructure đ„
Recently, DukaanÂź ditched the public cloud and started hosting their application on bare metal instances. This was a very counter-intuitive move and hence invited Subhash to the #AsliEngineering podcast to discuss it in-depth.
Throughout the podcast, Subhash gave live demonstrations of failovers, observability setups, infra scripts, trade-offs, and much more. Watch this podcast and learn from the legend himself.
watch it here - youtu.be/vFxQyZX84Ro
Do spread the word and let us know what intrigued you the most.
