Manish Gill
@manishgill.bsky.social
Engineering Manager @clickhouse.com | Databases | Kubernetes | http://parallelthoughts.xyz | https://berlinsystems.xyz/ | https://cal.com/manishgill | AI Research @TomTom
Delhi raised, Berlin based.
Delhi raised, Berlin based.
Reposted by Manish Gill
There was collaboration attempt between CMU, Tsinghua, Meta, CWI, Nvidia, Voltron, & SpiralDB. But then lawyers got involved and it fell apart. Everyone released their own format:
→ Meta Nimble: github.com/facebookincu...
→ CWI FastLanes: github.com/cwida/FastLa...
→ SpiralDB Vortex: vortex.dev
→ Meta Nimble: github.com/facebookincu...
→ CWI FastLanes: github.com/cwida/FastLa...
→ SpiralDB Vortex: vortex.dev
GitHub - facebookincubator/nimble: New file format for storage of large columnar datasets.
New file format for storage of large columnar datasets. - facebookincubator/nimble
github.com
October 1, 2025 at 1:49 PM
There was collaboration attempt between CMU, Tsinghua, Meta, CWI, Nvidia, Voltron, & SpiralDB. But then lawyers got involved and it fell apart. Everyone released their own format:
→ Meta Nimble: github.com/facebookincu...
→ CWI FastLanes: github.com/cwida/FastLa...
→ SpiralDB Vortex: vortex.dev
→ Meta Nimble: github.com/facebookincu...
→ CWI FastLanes: github.com/cwida/FastLa...
→ SpiralDB Vortex: vortex.dev
So if you have the depth + breadth, DM me.
May 26, 2025 at 3:27 PM
So if you have the depth + breadth, DM me.
But that's just an example.
The pattern I've usually seen is that DB developers can have *depth* in one area of expertise (think Storage, Query Processing, Consensus etc).
But for building a prod DB offering, depth isn't enough. You need significant breadth.
The pattern I've usually seen is that DB developers can have *depth* in one area of expertise (think Storage, Query Processing, Consensus etc).
But for building a prod DB offering, depth isn't enough. You need significant breadth.
May 26, 2025 at 3:26 PM
But that's just an example.
The pattern I've usually seen is that DB developers can have *depth* in one area of expertise (think Storage, Query Processing, Consensus etc).
But for building a prod DB offering, depth isn't enough. You need significant breadth.
The pattern I've usually seen is that DB developers can have *depth* in one area of expertise (think Storage, Query Processing, Consensus etc).
But for building a prod DB offering, depth isn't enough. You need significant breadth.
Of course, they might know a lot about Cloud Infrastructure -> all the way from AWS primitives down to what the hardware can support.
May 26, 2025 at 3:26 PM
Of course, they might know a lot about Cloud Infrastructure -> all the way from AWS primitives down to what the hardware can support.
Then there are outstanding applications - people who've got both the breadth and the depth.
They can:
Start from a log line ->
Think about DB operations in K8s ->
Reason the Kernel's allocation behavior ->
Impact on the DB's own memory allocation ->
Find a fix.
They can:
Start from a log line ->
Think about DB operations in K8s ->
Reason the Kernel's allocation behavior ->
Impact on the DB's own memory allocation ->
Find a fix.
May 26, 2025 at 3:26 PM
Then there are outstanding applications - people who've got both the breadth and the depth.
They can:
Start from a log line ->
Think about DB operations in K8s ->
Reason the Kernel's allocation behavior ->
Impact on the DB's own memory allocation ->
Find a fix.
They can:
Start from a log line ->
Think about DB operations in K8s ->
Reason the Kernel's allocation behavior ->
Impact on the DB's own memory allocation ->
Find a fix.
Can't hire a pure Infra person who cannot write code.
Can't hire a pure backend person who doesn't do infra - because usually there is a separate team to manage it for them.
Also can't hire a SQL expert who opens Snowflake and writes queries all day.
Overlap in ^ puts you ahead of the curve
Can't hire a pure backend person who doesn't do infra - because usually there is a separate team to manage it for them.
Also can't hire a SQL expert who opens Snowflake and writes queries all day.
Overlap in ^ puts you ahead of the curve
May 26, 2025 at 3:26 PM
Can't hire a pure Infra person who cannot write code.
Can't hire a pure backend person who doesn't do infra - because usually there is a separate team to manage it for them.
Also can't hire a SQL expert who opens Snowflake and writes queries all day.
Overlap in ^ puts you ahead of the curve
Can't hire a pure backend person who doesn't do infra - because usually there is a separate team to manage it for them.
Also can't hire a SQL expert who opens Snowflake and writes queries all day.
Overlap in ^ puts you ahead of the curve
If you're primarily a DevOps Engineer or an SRE, for many CVs, I don't even know if this person can code or not.
Usually, there is nothing in the CV that indicates application / backend development skills.
So can you build production grade systems?
Usually, there is nothing in the CV that indicates application / backend development skills.
So can you build production grade systems?
May 26, 2025 at 3:25 PM
If you're primarily a DevOps Engineer or an SRE, for many CVs, I don't even know if this person can code or not.
Usually, there is nothing in the CV that indicates application / backend development skills.
So can you build production grade systems?
Usually, there is nothing in the CV that indicates application / backend development skills.
So can you build production grade systems?
Conversely, if you're a k8s expert, but know nothing about the nuances of what it takes to manage databases, you will have an interesting learning curve ahead.
k8s itself is a vast ecosystem though, so the ability to deep dive into k8s control loops puts you ahead of the curve.
k8s itself is a vast ecosystem though, so the ability to deep dive into k8s control loops puts you ahead of the curve.
May 26, 2025 at 3:25 PM
Conversely, if you're a k8s expert, but know nothing about the nuances of what it takes to manage databases, you will have an interesting learning curve ahead.
k8s itself is a vast ecosystem though, so the ability to deep dive into k8s control loops puts you ahead of the curve.
k8s itself is a vast ecosystem though, so the ability to deep dive into k8s control loops puts you ahead of the curve.
3. Operationally, DBAs != Cloud Native Database Infra Devs. If you are used to database tuning to support production workloads in a few companies, you're off to a great start.
But can I trust you to build similar things on top of k8s?
But can I trust you to build similar things on top of k8s?
May 26, 2025 at 3:25 PM
3. Operationally, DBAs != Cloud Native Database Infra Devs. If you are used to database tuning to support production workloads in a few companies, you're off to a great start.
But can I trust you to build similar things on top of k8s?
But can I trust you to build similar things on top of k8s?
2. User != Operator != Developer, but the lines do get blurred.
The best DB developer I know is also the best user of said DB.
Operators don't have to be expert devs, but must be willing to get their hands dirty.
The best DB developer I know is also the best user of said DB.
Operators don't have to be expert devs, but must be willing to get their hands dirty.
May 26, 2025 at 3:24 PM
2. User != Operator != Developer, but the lines do get blurred.
The best DB developer I know is also the best user of said DB.
Operators don't have to be expert devs, but must be willing to get their hands dirty.
The best DB developer I know is also the best user of said DB.
Operators don't have to be expert devs, but must be willing to get their hands dirty.
www.foundationdb.org
February 21, 2025 at 4:07 PM