kunaldargan

Multimodal AI | MS(Research) @ IIT D | Computer Vision | ML Systems | Medical AI

[Xiangyi Li @xdotli]

before the age of slop

this is amazon internal guide on how to write

[codila @0xCodila]

Ex-Google Jeff Dean just released 1-hour lecture on full AI engineering: LLM → prompts → agent teams → graphs from 0% to 100%:

0% → 1:45 - LLM from scratch - that made Google
30% → 17:22 - how to actually use AI models
65% → 30:03 - prompt engineering
100% → 52:35 - one human coordinating 100 agents

this is 27 years of AI in Google compressed into one hour, by the person who lived it

watch it today - then read how to build graphs in the article below ↓

[Vivek Galatage @vivekgalatage]

Another great resource, the LLM Inference Handbook, from @Modular

https://t.co/rC1JzYqefh

[Puneet Patwari @system_monarch]

I'm a Principal engineer & I passed system design rounds of Amazon, Atlassian, Walmart, Saleforce, and Deliveroo.

Trust me, learning system is not hard. Start from these fundamental concepts:

1) Load Balancing: https://t.co/3jKCLiI6vl

2) CDN: https://t.co/dxzCmm9gAf

3) Caching: https://t.co/pRgn0FTPp2

4) Cache Invalidation: https://t.co/QrfRjJ57gd

5) Rate Limiting: https://t.co/LE5ECM2tGt

6) API Gateway: https://t.co/DgU8cBDUVr

7) CAP Theorem: https://t.co/a8WydnAIxd

8) Sharding: https://t.co/XQLU6eDriD

9) Replication: https://t.co/KuDkFH0fjx

10) Partitioning: https://t.co/3WXKeZLbLa

11) Queues: https://t.co/JchEoCcFmF

12) Microservices: https://t.co/aAQfM6AWMq

13) Microservices Vs Monoliths: https://t.co/bTaIIWkPU3

14) Fault Tolerance: https://t.co/qXNBoyOqYT

15) Database Scaling: https://t.co/D2lvPm1wkB

16) Service Discovery: https://t.co/z2DpwbJBVI

17) Consistency models: https://t.co/K2r3nMcCQu

18) Eventual Consistency: https://t.co/SWiz4ckIKR

19) Distributed Transactions: https://t.co/xqL7BTJxXn

20) Leader Election: https://t.co/ApNaYSnSFj

21) Horizontal vs Vertical Scaling: https://t.co/IFuEmzMfob

22) Back of the Envelope Estimation: https://t.co/7ntEmtVggQ

23) Idempotency, Data Latency & Finale: https://t.co/fNArLx4MrW

Let me know what you'd like me to cover, would love to help :)

[Yarchi @undefinedKi]

Airbnb just published how they run evals internally, and it reads like a job description for an AI engineer

Three layers. Programmatic checks first, an LLM judge second, humans last and only to calibrate the judge.

The numbers they work to: golden sets of 50 to 100 examples that have to include failures, judges calibrated to high 80s or 90s agreement with a human, measured with Cohen's kappa. 5% of live traffic sampled every day.

The admission that makes it real: roughly three quarters of their LLM-generated reference answers came out different on every labeling run. Their eval was measuring its own noise.

They got a full cycle from weeks down to a day, mostly by caching identical outputs and training tiny LoRA adapters.

Nobody in a job ad calls this eval engineering, but this is the work. Full breakdown in the article below.

Bookmark this

[Caglar @caglar_ee]

Video lectures, Stanford CS 329A Self-Improving AI Agents Autumn 2025, by Aakanksha Chowdhery & Azalia Mirhoseini
https://t.co/NRGbNwgA3Q
https://t.co/hOHDyWH0ha
.

[Ahmed Dahbi @dahbiahmed]

my personal site now has a face that follows your cursor and a 3D bookshelf you can pull books off and spin.

none of this was necessary...

[Eshant Garg @MindVestorIn]

July 2025. While everyone was chasing fancy & expensive tech stocks...

We asked our Mindvestor subscribers to look at Hero MotoCorp and Bajaj Auto.

No hype. No noise. Just no-brainer swing trades.

- Hero MotoCorp: 48% gains in 5 months. Booked.
- Bajaj Auto: Reinvested profits (as it had not started moving yet). Currently running at 28% upside and moving towards our target.

Full track record & detailed analysis available at the mindvestor website [Link in bio]

—-------------------------------------------------------

Now we've found something nobody is talking about. Again.

A company you've never heard of, but it's already in your daily life.

Right now, somewhere in India...

- A grandmother is getting an OTP to check her pension.
- A student is verifying his first UPI payment.
- A father is getting a fraud alert that just saved his savings.

None of them know the name of the company that made it happen. But it happened.

1 in 3 of those messages - the OTPs, the alerts, the notifications - travel through one invisible company.

In Indonesia:
- It just blocked 2 billion scams in 6 months.
- Protected 100 million people.
- Prevented $500M in fraud losses.

London Business School made it a case study.

And interestingly:
- Trading at 70% discount from its peak.
- Promoters and FIIs quietly increasing holdings.
- Available at half the industry PE multiple.

I probably shouldn't say this, as anything can happen in markets.

But some trades just look too damn obvious. And this valuation gap? It's hard to ignore.

I did the deep dive. Read it here 👇
https://t.co/rQ4HOcslCL

[kunaldargan @kunaldargan]

@alokbishoyi97 Hi this looks interesting, are there requirements for computer vision expertise ?

[Anika @AnikaSomaia]

if you’re also curious about robotics hardware and don’t know a lot (like me), i found the best place to start!!

this website has an interactive breakdown of literally every component in a humanoid: skeleton, motors, batteries, reducers, sensors, all the way to cost breakdowns and sourcing

you can click through real robots (like boston dynamics, apollo) and watch the spec sheets update live

also just a joy to use https://t.co/DU0djS1aml

[Dr.Sayajirao Gaikwad @DietDrsayajirao]

🫀 Key Cardiac Risk Markers You Shouldn’t Ignore
Heart attacks in 20s & 30s are becoming frighteningly common.
Fit on the outside ≠ Healthy on the inside.
Here’s how to investigate early & prevent silent cardiac risk 👇

🔴 1. Lipid Profile is just the beginning
Check more than just total cholesterol.
✅ LDL (bad cholesterol)
✅ HDL (good cholesterol)
✅ Triglycerides (TG)
✅ VLDL

🔍 Ideal:
• TG < 100
• HDL > 50
• LDL < 100
• TG/HDL ratio < 2

🔴 2. ApoB & ApoA1 – Better than LDL
🔸ApoB = Number of artery-clogging particles
🔸ApoA1 = Protective particles

✅ Ideal ApoB < 90
✅ ApoB/ApoA1 Ratio < 0.6
This is a stronger predictor of heart disease than LDL alone.

🔴 3. Lp(a) – The silent genetic risk
🔸Not tested in basic reports
🔸Genetically inherited
🔸Can increase heart attack risk even if other markers are normal

✅ Ideal Lp(a) < 30 mg/dL
📌 Test it once in life—even if you’re young.

🔴 4. hsCRP – Inflammation matters
🔸High-sensitivity C-Reactive Protein
🔸Marker of chronic low-grade inflammation

✅ Ideal < 1.0 mg/L
🔥 Chronic inflammation silently damages arteries.

🔴 5. Homocysteine – B12 & Heart Link
🔸Elevated in B12/folate deficiency
🔸Promotes clotting & damages blood vessels

✅ Ideal < 10 µmol/L
📌 If high, correct B12, folate, B6 levels.

🔴 6. Fasting Insulin & HOMA-IR
🔸Most missed test. Insulin resistance often starts years before diabetes.

✅ Fasting insulin < 6
✅ HOMA-IR < 1.5

📌 Even with normal sugar, high insulin = silent metabolic risk.

🔴 7. CAC Score (Calcium Score CT)
🔸Non-invasive scan to detect early plaque in coronary arteries
🔸Score 0 = very low risk
🔸Higher scores = hidden artery blockage

📌 Highly recommended if you have family history or are >35 with risk markers.

🧠 Prevention is not just avoiding junk food.
It’s about testing early and tracking the real markers of silent cardiac risk.
✅ Don’t wait for chest pain.
✅ Don’t assume fitness = health.
✅ Investigate > Guess
✅ Act early = Live longer
#HeartHealth #PreventHeartAttack #MetabolicHealth

[K2S @k2sbhai]

nvidia is casually giving you access to 5 frontier chinese AI models for free 😳

no credit card
no subscriptions
just one API key that unlocks everything

what you get for $0:

- DeepSeek V4 Flash for ultra-fast inference
- MiniMax M3 as a drop-in coding assistant
- Qwen3.5-397B for advanced reasoning tasks
- Kimi K2.6 for agentic workflows and long chains
- GLM 5.1 as a reliable everyday model

why this is huge:

> no paying separate subscriptions for different models
> no changing your existing workflows or tools
> no vendor lock-in since everything is OpenAI-compatible

getting started takes less than 2 minutes:

1. go to https://t.co/q50rSatNbb
2. sign up and verify your account
3. generate your nvapi key
4. set your base URL to https://t.co/92kkbFSf8D
5. pick any model and start building

supported models:

> minimaxai/minimax-m3
> qwen/qwen3.5-397b-a17b
> moonshotai/kimi-k2.6
> zhipuai/glm-5.1
> deepseek/deepseek-v4-flash

pro tip:

use DeepSeek V4 Flash for speed, Qwen for hard reasoning, Kimi for agents, and MiniMax as your daily coding companion

the best part?

one free key gives you access to 100+ models across NVIDIA's catalog

~40 requests per minute is more than enough for most developers and personal projects

5 frontier models that compete with GPT and Claude, all without spending a dollar

bookmark this and claim your free API key before the limits change 👀

[Alex Veremeyenko @alex_verem]

i'm obsessed with what's happening in AI reforestation right now

this Franco-Brazilian startup called MORFO took a patch of land in Brazil that was rock-hard and compacted from years of cattle farming. they replanted it using a single drone. months later the ground was covered in grass, bushes, and small trees. the land came back to life.

here's how the whole thing works.

1. drones scan the terrain with high-resolution cameras and sensors
2. AI analyzes the imagery alongside soil samples, moisture levels, slope, and surrounding vegetation
3. the system picks from a catalog of 300+ native species, deciding exactly which plants will thrive in which specific spot
4. the drone fires biodegradable seed pods packed with seeds, nutrients, and moisture at 180 capsules per minute
5. satellite and drone imagery monitors regrowth over time, with AI tracking vegetation cover and biodiversity
6. two people and one drone cover 50 hectares a day. a person planting by hand manages about one hectare.

and MORFO isn't alone. AirSeed in Australia drops 250,000 seed pods per day into bushfire-scarred koala habitat, replanting swamp mahogany that koalas depend on to survive. Flash Forest in Canada fires 50,000 pods daily into wildfire-destroyed boreal forest, planning the replanting alongside Cree Indigenous communities. re-green won Prince William's Earthshot Prize after planting 6 million seedlings across 30,000 hectares of Amazon and Atlantic Forest.

five companies across four continents built this same approach independently. nobody coordinated. the physics of the problem demanded it.

knowing which seeds belong in which soil used to require years of ecological fieldwork, manual planting crews, and budgets that made large-scale restoration nearly impossible. now two people with a drone and an AI model trained on local soil data can replant 50 hectares before lunch.

this is the AI work that'll still matter in 50 years.

[Vidit Gujrathi @viditchess]

Chess engines tell you the best move.

But grandmasters are human, they don’t always play it.

So I built "Kibitz": a human move predictor for chess broadcasts. I trained this model on my Nvidia RTX 5080.

Then I made it run as a business by itself.

A channel buys the overlay, Hermes onboards them, charges via @stripe test mode, runs the broadcast, narrates with @NVIDIAAI Nemotron, tracks inference cost, and books its own P&L.

I build. Hermes operates.

This is my demo and entry for the @NousResearch × @NVIDIAAI × @stripe Hermes Agent Accelerated Business Hackathon.

[Bilawal Sidhu @bilawalsidhu]

This tech maps the physical world in 3d and snaps it perfectly to my camera feed at 60 fps. Visual Positioning Systems (VPS) are the under hyped backbone of spatial computing. This is how we connect the world of bits &amp; atoms.

[Ashutosh Maheshwari @asmah2107]

Inference optimizations I’d study if I wanted sub-second LLM responses:

Bookmark this.

1.KV-Caching
2.Speculative Decoding
3.FlashAttention
4.PagedAttention
5.Batch Inference
6.Early Exit Decoding
7.Parallel Decoding
8.Mixed Precision Inference
9.Quantized Kernels
10.Tensor Parallelism
11.Pipeline Parallelism
12.Sequence Parallelism
13.Graph Optimization (ONNX, TensorRT)
14.Dynamic Batching
15.Memory Offloading
16.Streaming Generation

[Md Ismail Šojal 🕷️ @0x0SojalSec]

You can now build & run a complete real electric circuit simulator in full 3D,,,

This is insane:
- Full 3D circuit simulator
- Runs real Arduino code
- Accurate analog & digital simulation
- Blinking LEDs, sensors, motors, LCDs, displays everything and more

Powered by AVR8js & Three.js magic.
Everything runs smoothly right in your browser.
Perfect for learning, prototyping, and teaching electronics.

Real Arduino code running in the browser with full 3D circuits.

- https://t.co/4cWk8IIExQ

[Akshay 🚀 @akshay_pachaar]

https://t.co/5lb52h4oT1

[Codez @0xCodez]

Anthropic Managed Agents Lead:

"At Anthropic, >90% of our engineers are building with self-improving loops. In 4-6 months, it will be 100%.

my agentic loops can run for hours without spending hundreds of dollars."

in this 40-minute podcast, an Anthropic team lead reveals how to build effective agents from scratch.

Agent → harness → loops → memory = modern agent

This one video will replace 10 paid courses on vibe-coding.

Watch it today, then explore the same setup in the article below.

[SaaS to Agent @saastoagent]

This is big!
Our team has taken the obvious next (but tough to execute) step in AI-adoption: Transforming a live software system to an agent. Not a wrapper, not an additional chatbot.

Here's how you can do the same. Teams/devs in India working on agents will find this useful.