سید امیر حسینی

برنامه نویس، توسعه دهنده متن‌باز

[سید امیر حسینی @hemedani]

حس می‌کنم وضعیت اقتصادی که درست بشه، مردم میگن آخیش دیگه نمی‌خوایم یه ذره هم راجع به سیاست حرف بزنیم، میشن شهروند مطیع قانون، و این برا بچه شیعه اصلا خوب نیست

[سید امیر حسینی @hemedani]

@federikto این حماقت بی اندازه که علم رو با بروکراسی و آزمون و نمره سالها تحقیر کردن به زودی تموم میشه
بالاخره تا چند سال آینده هوش مصنوعی این فرایند معیوب و مریض رو بی معنی می‌کنه، کسایی سراغ علم میرن که عاشق‌اند!

[سید امیر حسینی @hemedani]

@raj_SDE1_FAANG Please also check my backend framework, it's very performant:
https://t.co/ZBfVZVpurR

Please check these articles also:
https://t.co/QGMOiUzP4A

https://t.co/K4kqJV2uBb

[سید امیر حسینی @hemedani]

@saloniiio Please also check my backend framework, it's very performant:
https://t.co/HLLVUB5KHl

Please check these articles also:
https://t.co/QGMOiUzP4A

https://t.co/K4kqJV2uBb

[سید امیر حسینی @hemedani]

@uday_devops Please check my backend framework, it's very performant:
https://t.co/HLLVUB5KHl

Also read this article:
https://t.co/QGMOiUzP4A
And this
https://t.co/K4kqJV2uBb

[سید امیر حسینی @hemedani]

@cinamarina Hi, I'm working on Lesan: https://t.co/HLLVUB5KHl
Read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/168Fmccad4
the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb
Sorry for saying this here, I just wanted to introduce

[سید امیر حسینی @hemedani]

@JeenaKrish91074 Hi, please read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/168Fmccad4
the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb
Sorry for saying this here, I just wanted to introduce Lesan: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@vikrambuilds Hi, please read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/QGMOiUzP4A
Why the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb

Sorry for saying this here, I just wanted to introduce Lesan: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@thepoonam0914 Hi, please read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/QGMOiUzP4A
Why the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb

Sorry for saying this here, I just wanted to introduce Lesan: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@prajwalraj23 Hi, please read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/QGMOiUzP4A
Why the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb

Sorry for saying this here, I just wanted to introduce Lesan: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@senodev_ Hi, please read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/QGMOiUzP4A
Why the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb

Sorry for saying this here, I just wanted to introduce Lesan: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@the_codewala Hi, please read these articles:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/QGMOiUzP4A
Why the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb

Sorry for saying this here, I just wanted to introduce Lesan: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@ayobuilds @X Hi, please read these article and give me feedback:
Embedding relations turns ~2.5M-doc reads into ~25K: https://t.co/QGMOiUzP4A
Why the writes aren't a nightmare — the nature of data: https://t.co/K4kqJV2uBb
Open source: https://t.co/HLLVUB5KHl

[سید امیر حسینی @hemedani]

@jiya_3063 Related idea we explored: client-driven projections over a single POST endpoint — clients pick the fields and depth (like GraphQL) but there's no query language or schema server to maintain. The read-performance side of that: https://t.co/QGMOiUzP4A

[سید امیر حسینی @hemedani]

@ekemini58110 We hit this building an ODM for MongoDB — the nested-find / $lookup path fetched ~2.5M docs for one dashboard screen. Embedding with automatic sync got the same screen to ~25K docs, one query. Details + reproducible benchmarks: https://t.co/QGMOiUzP4A

[سید امیر حسینی @hemedani]

@SRProgramist We hit this exact wall. Nested find/$lookup fetched ~2.5M docs for one dashboard. Embedding + auto-sync dropped it to ~25K, single query.

Write-up + reproducible numbers: https://t.co/QGMOiUzP4A

@SRProgramist @MongoDB

[سید امیر حسینی @hemedani]

@iamrishit8 We hit this building an ODM for MongoDB — nested-find / $lookup path fetched ~2.5M docs for one screen. Embedding with automatic sync got the same screen to ~25K docs, one query.

Details + benchmarks: https://t.co/QGMOiUzP4A

@iamrishit8 @MongoDB

[سید امیر حسینی @hemedani]

@Sakshamsri744 N+1 is exactly the bug this solves structurally.

By embedding pure snapshots of the relation inside the parent (kept in sync on write), the deep traversal becomes one query. We went from ~2.5M docs to ~25K on a dashboard screen.

Benchmarks: https://t.co/QGMOiUzP4A

@MongoDB

[سید امیر حسینی @hemedani]

@ishankaru This is the tradeoff that scares everyone away from embedding — and the reason most teams go back to joins.

We made embedding safe: keep the sync bounded with limits/excludes so the expensive writes stay rare.

If useful: https://t.co/K4kqJV2uBb

@ishankaru @MongoDB

[سید امیر حسینی @hemedani]

4/ The framework doesn't just tolerate that — it nudges you to the better model. Full write-up: https://t.co/K4kqJV2uBb

[سید امیر حسینی @hemedani]

1/ "Embedding duplicates data — every write gets expensive." Fair objection. The answer lives in the nature of data: the writes that are expensive to maintain are exactly the ones that happen least often.

[سید امیر حسینی @hemedani]

3/ The scary case (population changing every 5s → 1B+ embedded copies): either `exclude` the volatile field, or promote it to its own model. "Every frequently-changing field can become a relationship."