Joined March 2021
393 Photos and videos
Update: Working on a distributed AI workflow runtime.
1
35
I think I should build in public. Anyway, I am going to make it BSL licensed.
33
Just learned about how Agentic AI and AI pipelines work, or I should say how Context Engineering works.... Damn. Most of it is just normal engineering.
23
Am I the only one who likes big monorepos?
38
Most of Computer Science and Software development is just common sense. And that's the reason why LLMs running on tokens are so good at it.
45
Fuck this Shit 🀯 Just got into a platform engineering rabbit hole on behalf of small-to-mid sized startup teams. That are struggling with building and deploying their code while arguing about monorepo vs multirepo setups wasting time and money on build pipelines. Will share.
1
1
65
How does your Company or team handle code? Not features, not bugs, not devops, nothing of that. Just pure code and it's scale. I think I am talking about platform engineering.
1
1
41
It's Learning for me. Can't beat the market by being stagnant or stupid.
Founders, how are you spending your weekend? - building - chilling - marketing
1
1
67
Thinking of starting one such OSS project myself, by forking diceDB. @arpit_bhayani πŸ˜‰
Unemployed SWE: Dozens of OSS projects on GitHub. Employed SWE: Dozens of take-home interview tasks stored locally.
1
79
Just learned, there's a tool by Google called Copybara, & it solves the problem of everything inside monorepos being visible to even people concerned with just a single submodule or project. Always, thought about how companies managed monorepo security. & Copybara is the answer.
30
The unspoken path to being a great engineer or a great product manager is knowing & having experience of the other side too. A PM who doesn't know how a feature is actually built in dev side, is just as lost and frustrated as a dev who doesn't understand business requirements.
1
27
We know that a database has a storage engine to work through the data structures their layouts & generic data manipulation But storage engines in turn, have several components that handle the underlying complexity & responsibility Typically a storage engine has 5 components πŸ‘‡
2
1
1
218
3. Access Manager - This component manages the data structures and their storage and retrieval as well as their organization on disk 4. Buffer Manager - Not everything can be done on disk, and that's why you need RAM. And this component talks to RAM and manages page buffers 2/3
1
1
45
Last one - 5. Recovery Manager - All the operation logs get stored somewhere as a side effect of all the above components, and this one manages just that. It ensures the database can fall back to last consistent state before things got wrong. 3/3
1
21
How many of you remember what OLTP, OLAP and HTAP databases are? For my forgetful folks, here's the breakdown πŸ‘‡ 0/2
1
48
OLTP - Online Transaction Processing DB. The most common type of databases in use that handle queries concurrently and can manage predefined and shot lived requests OLAP - Online Analytical Processing DB. These manage heavy complex and long queries for analytical, purposes 1/2
1
41
HTAP - Hybrid Transaction and Analytical Processing DB. As the name suggests, these are a weird mix and combination of OLTP and OLAP paradigms of database design. 2/2
24
Here's my list for top 5 reasons why startups fail: 1. Not having clear plan and goal. 2. Not knowing your customer. 3. Not knowing what to build. 4. Not knowing when to stop building. 5. Not knowing when to give up.
29