SJ

242 posts

SJ banner
SJ

SJ

@SJcorrupt

Living being | AI/ML • Cloud • Mlops • Backend

Earth เข้าร่วม Mayıs 2026
100 กำลังติดตาม89 ผู้ติดตาม
ทวีตที่ปักหมุด
SJ
SJ@SJcorrupt·
Day 22 (2) of ML • Started Ridge Regression. • Learnt about Ridge Regularization (L2) and worked on a toy dataset. • Revised few previous concepts: > Bias and Variance > Bias and Variance Tradeoff. • Even after revising these, there's few things I learnt in BVT is that to manage Overfitting in a ML model there's mainly three methods: > Regularization > Bagging > Boosting • Also got to know there's mainly three Regularization techniques: > Ridge (L2) > Lasso (L1) > Elastic Net • Still an Average day #buildinpublic
SJ tweet mediaSJ tweet mediaSJ tweet media
SJ@SJcorrupt

Day 21 (1) of ML (2.0) • Done with All the Linear regression concepts today. • 3rd pic represents the score difference between, before using Polynomial Regression and After using Polynomial Regression. • Tomorrow gonna continue the other Regression Algorithms. #buildinpublic

English
1
0
4
31
KARAN JAIN
KARAN JAIN@Karanjain2311·
Life update : Doing intern at @IITHyderabad under Prof. Shibam Gosh
KARAN JAIN tweet media
English
7
0
67
2.2K
SJ
SJ@SJcorrupt·
India can build fabs with money, but building talent is the real challenge. Projects like HackerFab IITB are creating the hands-on semiconductor experience our ecosystem needs.
Caleb@caleb_friesen

India is about to face a MAJOR semiconductor bottleneck. The Government of India has approved ~13 semiconductor projects under the India Semiconductor Mission, across 7 states. Three of these are full/compound fabs. Things are ramping up FAST, with ISM supported by an incentive framework of ₹76,000 crore. But one massive question mark remains: where is the talent going to come from? The money is there. The fabs are going to be there soon. But what about the many thousands of skilled technicians required to run these semiconductor fabrication plants? Much of the knowledge in this industry is tightly-guarded trade secrets kept under lock and key by the nations that lead global semiconductor production. One way India can quickly close this knowledge gap is by ensuring that young people across the country are learning how to fabricate semiconductors from first principles. Ideally at the university level if not earlier. But because this is an entirely new industry segment in India, most of the country’s top colleges haven't caught up. Semiconductor fabrication is not accessible to Indian students. Until the Graduate or PhD level, most students never even get to touch a silicon wafer. A group of 15 students at IIT Bombay wants to change this. 10 months ago they launched the HackerFab at IIT Bombay. So far, they’ve raised ₹30 lakh to built DIY machines like a DLP-based lithography machine, a tube furnace to oxidise silicon, and a DC plasma sputter. They realised that existing institutions weren’t going to give them the early education they needed to develop REAL chip fabrication experience, so they took up the challenge themselves and created everything from scratch. HackerFab IITB is one of the most important developments in India’s semiconductor story, not just because the students passing through this programme will become leaders in India’s future semiconductor industry, but because they’re open-sourcing the India-specific recipes they’ve developed to build their machines and processes. They’re doing this so that other Indian colleges can replicate their work. No more gatekeeping. This movement started at IIT Bombay, but it will spread to other Indian colleges soon. As a result, India will see young people graduating from college with practical semiconductor fabrication experience first the very first time.

English
0
1
5
106
SJ
SJ@SJcorrupt·
@dark_coderz Congrats man!! Crazy growth 📈
English
0
0
0
19
Dark Coder
Dark Coder@dark_coderz·
Can’t believe it! 😧 I reached 2060 followers only in just 40 days… it feels unreal.. 1160+ verfied follwers (58%) 8+ Articles posted Thank you everyone 🙏🏻
Dark Coder tweet media
English
65
0
75
1.9K
SJ
SJ@SJcorrupt·
@atulit_gaur Another idea: Dating App for Cab, Bike and Food delivery Riders :D
English
0
0
1
46
atulit
atulit@atulit_gaur·
startup idea: am app where cab drivers can talk to each other to avoid being bored
English
7
1
25
874
SJ
SJ@SJcorrupt·
@Siwe_e Will do Thanks 🤝
English
0
0
1
8
SJ
SJ@SJcorrupt·
Day 21 (1) of ML • Finally Starting with this series again. • Revised few Gradient Descent concepts today. • Practiced Linear regression and Multiple Linear Regression code from scratch. • Gonna Finish Linear Regression by tonight. #buildinpublic
SJ tweet mediaSJ tweet mediaSJ tweet media
SJ@SJcorrupt

Day 20 of ML • Started Gradient Descent. • Worked on how Gradient Descent works step-by-step. • Working on building a class to find intercept. • Gonna learn more for few days on GD. #buildinpublic

English
1
0
15
516
SJ
SJ@SJcorrupt·
@zarath_dev Means a lot mate🙌
English
0
0
1
15
SJ
SJ@SJcorrupt·
Day 21 (1) of ML (2.0) • Done with All the Linear regression concepts today. • 3rd pic represents the score difference between, before using Polynomial Regression and After using Polynomial Regression. • Tomorrow gonna continue the other Regression Algorithms. #buildinpublic
SJ tweet mediaSJ tweet mediaSJ tweet mediaSJ tweet media
SJ@SJcorrupt

Day 21 (1) of ML • Finally Starting with this series again. • Revised few Gradient Descent concepts today. • Practiced Linear regression and Multiple Linear Regression code from scratch. • Gonna Finish Linear Regression by tonight. #buildinpublic

English
3
0
12
266
Tanvi Gawade
Tanvi Gawade@TanviGawade13·
Day 1 -Learned the basics of how the Internet works behind the scenes. -Refreshed my JS skills by building a To-Do List. -Revised Time Complexity and Space Complexity concepts. -Learnt API routing with Node.js and Express.js. -Studied key Statistics concepts for my AI/ML journey.
English
1
0
2
61
SJ
SJ@SJcorrupt·
@yellow_void22 Good work for today!! You don't have summer holidays in college or what??
English
1
0
1
25
Maestro
Maestro@mohitbareja999·
💼Day 24 – #SDESheetChallenge One of the easiest days yet! Took me 10 minutes to solve! Spent the rest of my time in codechef contest. Missed the 2-star rating by just 16 points! 🌟 🟢Implement Stack, Queue using Arrays 🟢Implement Stack using Queue and vice versa
Maestro tweet mediaMaestro tweet mediaMaestro tweet mediaMaestro tweet media
English
3
0
20
317
SJ
SJ@SJcorrupt·
@zarath_dev 75% soon Congrats 🤝
English
1
0
2
29
SJ
SJ@SJcorrupt·
@levidiamode Does PC or Laptop matters for learning GPU programming? I have heard it requires atleast a good processor to manage tasks.
English
0
0
0
66
levi
levi@levidiamode·
172/365 of GPU Programming My agents went a bit rogue while I was sleeping and decided to sneak in a few reward hacks here and there, so spending some time today going through my recent submissions and deleting the faulty ones that made it past the leaderboard checks. Interestingly, I think the more exhaustive the search is in terms of legitimate methods tried the more creative the models seem to get with their reward hack approaches. Maybe the real treasures lie not in the optimizations but the reward hacks you learn along the way.
levi tweet media
levi@levidiamode

171/365 of GPU Programming Finally got below 2ms on the GPU Mode QR challenge. Still quite far away from CUDA colonel @blelbach, so I'm excited to read everyones' writeups once the competition is over. I especially wonder how #1 is utilizing NVFP4 (or is the submission name a false flag?). Every lower precision attempt has not panned out for me so far unfortunately. Very curious where the top 3 will converge in a week from now... If you're still debating whether to participate, would highly recommend! You learn so much from just trying stuff and being in the discord. And @modal gives you $30 in monthly compute for free. Maybe we can get a bit more compute for the next challenge if we all ask @charles_irl nicely 😁😁

English
2
1
56
2.7K
SJ
SJ@SJcorrupt·
@parmita_ilya Shift few things for the weekends rather than stacking on a normal day.
English
1
0
3
73
Parmita Rajthala
Parmita Rajthala@parmita_ilya·
Time Management Issue!! I was out almost the whole day for some work. By the time I got home,I was exhausted, but I still had my React learning session and LeetCode to finish. It honestly felt overwhelming. How do you guys balance a busy day with consistent learning? Any tips?
English
11
0
30
564
SJ
SJ@SJcorrupt·
@karanbuidls 50 Lpa is a great target Work hard🤝💪
English
1
0
2
60
Karanjot Singh
Karanjot Singh@karanbuidls·
`Road to 50 LPA` #2 Dived into Agentic AI infrastructure & execution flow today Today's learning: - Shortcomings of langchain - 6 types of context - Graph definition, Execution & State mngmnt layer in LangGraph *Btw do search about Google Pregel - BSP & isolated super-step :)
Karanjot Singh tweet media
English
13
1
77
1.3K
SJ
SJ@SJcorrupt·
Day 15 of Backend Spent the day revising the last 14 days of learning. • FastAPI: Pydantic, HTTP methods, validators, computed fields, serialization • Docker: Dockerfile, images, containers, registry, Docker Hub • PostgreSQL: CRUD, datatypes, constraints #buildinpublic
SJ tweet media
SJ@SJcorrupt

Day 14 of Backend • Done with PostgreSQL (enough to explain things). • Worked with VIEWS. • Worked with HAVING clause. • Worked on few user defined functions. • Worked with WINDOW function. • Learnt about CTE (Common table expression) and Triggers. #buildinpublic

English
7
0
21
500