Sashi
@t_sashidhar
https://jobq.run/ - Run background jobs at any scale with no extra infra. Past: @UberEng, @Walmart, @Olacabs.
I'm building a language-agnostic background job server that's durable, scalable and reliable. Cherry on the cake it doesn't require any additional infra like Redis or a database or pub/sub for it's operation. #backgroundjobs
New post: "Make Worse Software, Slower" This is not satire. I argue for: – Rejecting functional programming – Using ORMs to perfectly hide impedance mismatches – Skipping event sourcing – Other proven best practices Link in replies.
‘Hindi Imposition’ is not what you think. The real victims? They can’t even speak up. #MustRead 🚨 The debate you're seeing all over India about ‘Hindi’ isn't the real pain point. In fact, it's distracting from what actually matters. From a tribal child in a forest village to a…
OOP is not a methodology. It is a style of programming that prioritizes the separation of concerns through the management of source code dependencies.
AI is not an excuse to relax your discipline. On the contrary, effective use of AI requires a greater reliance on discipline and intense scrutiny. TANSTAAFL!
We are in the mainframe era of AI Computing!
Writing text back and forth with an LLM is like we're all the way back to the era of command terminals. The "correct" output is a lot closer to custom web apps written just for your query, information laid out spatially, multimodal, interactive, etc. Will take some time.
Morning bathrobe rant at the cabin: performance.
Need document data model ? Just use a Map of Maps. Need columnar analytics? Use a Map of Sorted Maps. Want key-value lookups? That's just a simple Map. The power is in the shape you define.
Traditional backend: • You have an application layer • You have a database layer You’re always trying to get them to agree. The result? Impedance mismatch, fragile ETL pipelines, painful evolution.