Harshit Bansal (@harshitcodes) header
Harshit Bansal (@harshitcodes) profile picture

Tech x Finance • engineering @avenueticketHQ • building https://teapawt.net/t.co/eYDYLNZ3PU

del/bom
Joined August 2024
Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)Photo by Harshit Bansal (@harshitcodes)
Harshit Bansal retweeted
India mein koi 5 minutes number Google kar le to doodh ka doodh paani ka paani ho jaye. RBI surplus to Govt : ₹2.87 Lakh Crores (US$30 Bn) Total listed bank profits : ₹4.11 Lakh Crores (US$42 Bn) NPCI - which runs UPI pre tax ‘surplus’ : ₹ 1,888 crores (US$200 Mn) To nuksaan kis ka ho raha hai UPI se aur kaunsi subsidy de rahi hai Govt UPI pe jo chubh rahi hai? Cost of running ATMs and cash logistics in India is ₹30,500 crores (US$ 3.0 Bn). If you want to optimise shut ATMs and promote UPI instead. Any levy on UPI is just tax collection. UPI is the one scientific achievement of India everyone acknowledges par ab tax ki bali chadhegi.
UPI Charges Row | Ashneer Grover Exclusive If I transfer ₹1 lakh to my friend through UPI, or I pay ₹3,000 using a QR code, what is the cost of running the system? The cost is the same. So, if I transfer ₹1 lakh to my friend, nobody is making money from it. But if I make a payment of ₹2,500 to a small shopkeeper then the government wants to put charges, what’s the logic? — @Ashneer_Grover speaks to @shreyadhoundial
1,874
14,982
600
59,335
4,400,504
Harshit Bansal retweeted
Our only unfair advantage left is wondering. Not knowing. Not having the answer. Just being curious enough to ask, "What if?" Machines are getting better at knowing. They are getting better at finding answers, connecting ideas, and even producing things we once needed years to learn. Perhaps our advantage is not in knowing more, but in knowing what is worth wondering about.
25
38
2
626
26,226
Harshit Bansal retweeted
My mentor at Rubrik was a super talented chinese principal engineer. One golden piece of advice he gave me was: “Learn to build a strong opinion on things. No strong engineer you look upto just says yes to everything” This has been a game changer. Learn to reason and be vocal with your reasoning. Its okay to be wrong, you can always correct your understanding. You are hired to reason out decisions and build great products, not to bootlick.
21
73
6
1,424
41,898
Harshit Bansal retweeted
been using AI to write code for almost a year now and i still keep seeing the same issues, even with frontier models like Fable5, Opus 4.8 and GPT-5.5 1) if "A" dependency already does it, AI installs "B" anyway. 2) 30 lines of logic somehow turns into a 200 line function. 3) a simple if else becomes 50 lines of abstractions which isnt needed anyways. worked with quite a few teammates too and whenever i asked "why was this change even needed?" or "why this PR?", there was almost always silence way too many people are relying on AI without thinking about the most important part of coding: knowing what not to write.
16
5
111
5,810
You focus on the 1 little thing they got wrong, They focus on 100 things they had to get right and a 1000 things that were nice-to-haves, but deprioritized
2
2
59
3,837
Harshit Bansal retweeted
Replying to @mannupaaji
I started with development and when I tried doing DSA, I realised that it isn’t for me so I stuck with development. Now I am doing a job with my college where I built a product from scratch with 3 team members. It’s not mine but that product is going to hit 10 million ARR in the next 1-2 months. It’s really satisfying to see this growth and seeing the product which I built getting into scale This wouldn’t be possible if I was just grinding DSA
5
4
42
1,886
Harshit Bansal retweeted
Had a detailed engineering discussion with my team after so long. It felt really good to use brain instead of AI
2
2
1
21
1,417
Harshit Bansal retweeted
How to win at software engineering, in 47 words: Ship, test, refactor, document, and review. Build, automate, measure, debug, and deploy. Read code, write daily, delete often, question specs, and ignore the hype. Make these habits. Say no to everything else. Avoid tech debt, scope creep, ego, burnout, shortcuts, and pointless meetings. Relax. Mastery is inevitable.
37
6
50
2,131
Harshit Bansal retweeted
AIN'T NO WAY THE JOB MARKET GOT MFS DOING THIS...
1
7
463
Harshit Bansal retweeted
THIS!!!!!
105
29,304
798
176,667
3,468,571
I love the `react-hotkeys-hook` package for keyboard bindings.
10
17
2
409
22,457
If your app does not have bugs in the initial version, you've shipped too late.
1
1
4
178
Harshit Bansal retweeted
- you write some scripts to set up ec2 and nginx - spin up an ec2 instance - configure standard security groups - ssh into your machine - clone the backend repo (no monorepo yet) - setup .env - run a few setup scripts - spin up some docker containers - run out of ram very quickly - optimize the dockerfile for production use - spin up a new ec2 instance - spin up the docker containers again - run df -h - discover 99% of storage is already used - spin up another new instance, optimize again - run deployment scripts and set up nginx and cron jobs - postgres and other containers start working - but there's no response from curl localhost:3000 - check docker logs - find issues in the production docker-compose - resolve the issues, push, pull, retry - try a few more times - finally figure it out, things start working, server is up - point your subdomain to the host ip - set up ci/cd - your own tests start failing, pipelines are broken - 30 commits and 10 prs later, you figure out all the issues, each action takes 3 minutes btw, and 5 mins to test - finally, one github actions pipeline succeeds - no certificate, run certbot manually, generate a self-managed ssl - pg_dump to copy data from the old database - run into some username and password issues while restoring the new db - dump as clean sql, expose port 5432, restore everything back, seed admin - frontend integration is delayed, 443 wasn't exposed - prod db out of sync - back up db again and restore to prod db once more - write system monitoring scripts, protect the main branch - cors issue on the status pages; solved status page is up - use the old server as a staging decoy for testing - rate limiting not working, ssh again into the server to monitor - again wait two second after typing each character in ssh to get response - write more scripts for rate limit testing - split the rate-limiting middleware into two, general and strict - create a service to sync production db with staging db - set up logging and backup facilities on the server "yeah bro, just move ec2"
40
22
6
523
37,820
Harshit Bansal retweeted
ECE people are better software developers than CSE people
25
4
2
135
6,140
Harshit Bansal retweeted
my first mechanical keyboard
49
2
1
456
16,340
You aren't the average of your 5 closest friends. You're the average of 5 people you allow to influence you.
12
25
3
310
10,742
If you are planning to learn the "MERN" stack, please stop, learn the 'R' i.e React and learn something else for the backend part like Java (Springboot) or Golang (actual requirement of the industry). I still don't understand how this stack became so popular.
53
46
7
948
69,552
X is tough
3
2
1
19
934
Harshit Bansal retweeted
2
1
11
501
Harshit Bansal retweeted
😌😌
44
179
32
1,945
91,026