Rachit

engineer. @atlanhq
building @crosmoslabs, prev @localhosthq
absorb all you can

[Rachit @rachitcodes]

https://t.co/o0hZbjO57S

[Rachit @rachitcodes]

@traceliteai that's a good addition. i was thinking about provenance in terms of maintaining the connection between the original trace and the derived eval, but validating whether the derived case actually preserves the original failure conditions is another important step.

[Rachit @rachitcodes]

https://t.co/o0hZbjO57S

[Rachit @rachitcodes]

@dipxsyy thank you :)

[Rachit @rachitcodes]

have some more bangers incoming 👽

[Rachit @rachitcodes]

https://t.co/o0hZbjO57S

[Rachit @rachitcodes]

@ronish hell nah😭

[Rachit @rachitcodes]

@sharpeye_wnl goated

[Rachit @rachitcodes]

@bilal2vec ball

[Rachit @rachitcodes]

@NathanFlurry herdr is gpui

[Atlan @AtlanHQ]

We’re excited to launch Atlan Trident 🔱

The shared context workspace for humans and agents.

One place for every skill, agent, metric and ontology.
Shared across your team, in the tools they already use.
Better with every run.

[Rachit @rachitcodes]

@AtlanHQ 🔱🔱

[Rachit @rachitcodes]

@vaxryy look out for her asking for your hyprland config

[Rachit @rachitcodes]

@Yash_s33244 thank you :)

[Rachit @rachitcodes]

@get_soju @dokidek_ lesgoo

[Rachit @rachitcodes]

@dipxsyy @ethantroy yoo lesgoo king

[Rachit @rachitcodes]

@ronish my goat 🫶

[Rachit @rachitcodes]

@pass_at_k thanks alot :)

[Rachit @rachitcodes]

@pass_at_k yes, this is an important pointer. having a separate tune/test dataset split helps ensure we're measuring actual improvements rather than just optimizing for the evaluation set. refreshing the holdout with new production failures is equally important.

[Rachit @rachitcodes]

evals evals evals evals

[Rachit @rachitcodes]

https://t.co/o0hZbjO57S

[Rachit @rachitcodes]

@benhuds @PrathmeshPatel_ @rishigb this is awesomee 🚀