Sparsh Nagpal 🧑🏻‍💻

26 | AI Products Builder

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

https://t.co/hf8xxqynRg

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

Built Claude Code skills that simulate advisory conversations with famous technologists, founders, scientists, and thinkers. Each board member is a deployable skill built from primary sources, verified quotes, documented decisions, and evidenced mental models.

Link below:

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

Subagents cost token, use claude to build. Let free subagents look into research data. built a tool, “nim-pi-research-agent”, a FastMCP server that delegates research to Pi subagents on free NVIDIA APIs.

https://t.co/wXri1knHVF

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

Org LLM bills explode from overlapping queries across users, yet no tool handles safe cross-user cache reuse end-to-end.

I built PRISM-Cache: a Tier 0-4 caching pipeline for safe prompt sharing. Evals show 98% hits & zero data leakage.

Code: https://t.co/G4ixpE3Uba

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

Current LLM tool setups break at scale: they're too model-heavy (massive token overhead) or shell-heavy (zero governance)

I built GAX: a governed execution runtime for AI agents. It features lazy tool discovery, capability-scoped execution & audit logs.✌️
https://t.co/9IXKd3vy6b

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

Memorable: The first memory system that combines:
✅ Research-backed techniques (like Mem0)
✅ Zero-code integration (like Memori)
✅ Optional graph db (like Supermemory)

Use any SQL DB. 100+ LLMs. 2 lines of code.

🚀 memorable-ai on PyPI
💻 0sparsh2/memorable-ai on Github

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

Stable Diffusion was released less than a year back
They want people in GenAI with 3+ years of experience :)

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@gdb Going to witness a defining moment in the AI industry

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

#chatgpt #opensource #thread #aicommunity

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

ChatGPT now open source?

Not exactly but there's surely an equivalent alternative emerging

PaLM + RLHF (Reinforcement learning with human feedback

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

https://t.co/QPlmLAC5vr

Check out the GitHub repo here

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@sakshirambhia https://t.co/6ArqNrjJRS

Thanks

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@svpino A Machine Learning Engineer here with just a bachelor's degree working fresh out of college. Happy to help

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@0xSmit @kalpitveerwal Saying as a recent pass out of '22 CS engineering batch from a Non-bits Non-iit Non-nit college,
It's very much possible to get a 50 lakh package in India as a fresher. And you don't need to be coding since 12, at all.
I myself have witnessed so my people reach there.

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@anuragstwt Package not found

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@mfers_0 Hey,
Check your DM, just sent ✌🏼

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@elina68208230 @Globalnft02 I'll do that, thanks!

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@mfers_0 You can also access my opensea profile through the link I shared

[Sparsh Nagpal 🧑🏻‍💻 @TechSparshyyy]

@mfers_0 You don't have your DM option open