Meghraj Deshmukh

60 posts

Meghraj Deshmukh banner
Meghraj Deshmukh

Meghraj Deshmukh

@itsdeshmukh

Software Craftsman -🧑‍💻 System Design

Katılım Şubat 2025
330 Takip Edilen4 Takipçiler
Meghraj Deshmukh
Meghraj Deshmukh@itsdeshmukh·
@LearnersBucket I have been lately experimenting with multiple LLM , One thing is clear ,any backend fullstack guy who has a fundamental understanding of framework and UI can independently work on any frontend app, rarely I have seen the issue with code , so no depth on FE anymore
English
0
0
1
166
Prashant Yadav
Prashant Yadav@LearnersBucket·
My current workflow looks like: - Create a plan from PRD and Figma (basically an advanced ERD) in Claude code. - Review the plan within team - Use this plan to create JIRA tasks with desc - Then use the plan + JIRA tasks for dev in Cursor Now I am following SDLC
English
2
0
44
2.8K
prateek
prateek@agent_wrapper·
join my team for the upcoming 𝗖𝗼𝗱𝗲𝘅 𝗵𝗮𝗰𝗸𝗮𝘁𝗵𝗼𝗻 if you have gotten accepted for the Codex hackathon already, and don't have a team yet, here is my pitch we are the people behind 𝗔𝗴𝗲𝗻𝘁 𝗢𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗼𝗿 (@aoagents launched Feb 22) Agent Orchestrator is the most popular agentic coding orchestrator on github right now, and the most pluggable way to run and manage coding agent swarms today in about 7 weeks since launch, we are already sitting at a cool ⭐ 𝟲.𝟮𝗞 𝗚𝗶𝘁𝗵𝘂𝗯 𝘀𝘁𝗮𝗿𝘀 we are also the first result on google when you search "Agent Orchestrator" (all organic) since we launched AO, the biggest names in AI have started building their own orchestration layers 𝗖𝗼𝗴𝗻𝗶𝘁𝗶𝗼𝗻 shipped "Manage Devins" — one Devin can now break down tasks and delegate to a team of managed Devins running in parallel. 𝗖𝘂𝗿𝘀𝗼𝗿 launched 3.0 with an Agents Window to run many agents in parallel across repos and environments. 𝗚𝗶𝘁𝗛𝘂𝗯 shipped Copilot Cloud Agent. 𝗔𝗪𝗦 𝗟𝗮𝗯𝘀 built their own CLI Agent Orchestrator. from YC startups to the biggest AI labs, everyone is converging on orchestration as the default architecture for how coding agents work. we saw this coming and shipped it open source 7 weeks ago. and the funny thing is, Agent Orchestrator is just a side project imagine what we can build when we go all-in for a full day hackathon if you're in Bengaluru, got accepted, and want to team up, DM me. let's build something insane.
prateek tweet media
English
12
3
77
8K
neural nets.
neural nets.@cneuralnetwork·
Everyone says Projects are important but no one tells what projects are the best tell me something about yourself and I'll give you a project idea
English
144
16
720
36.6K
Meghraj Deshmukh
Meghraj Deshmukh@itsdeshmukh·
@championswimmer Hey @championswimmer, what's the best way to make a brand stand out by doing projects? What kind of projects are still good for getting an interview if you're trying to build a brand? 4+ year experience
English
0
0
0
46
Arnav Gupta
Arnav Gupta@championswimmer·
I still keep asking... who is using "swarms of agents" and getting any thing really done? If you are actually doing it - then you have the "skill issue", honestly. Here two subagents which should have been serial, started in parallel instead, and the 'reviewer' agent says it is completed it's task even before the 'worker' agent has even written all the code No this doesn't always happen. It in fact happens less often than it works fine. But still like 1 out of 5 time of they work like this, how am I supposed to 'spawn up a team of agents' and sleep peacefully knowing this is how they might be working? And the bigger 'agent teams' you spawn, the larger volumes of code they write, and less you can manually read later to see what they wrote.
Arnav Gupta tweet media
English
21
4
131
13K
Arpit Bhayani
Arpit Bhayani@arpit_bhayani·
What would it take to implement a database wire protocol, and how should you think about it? In the last video, we discussed what a wire protocol is and why we need it to talk to databases. Today, I just published the fourth video in my Redis Internals series, where we implement it. We literally implement RESP (Redis Serialization Protocol), and this enables us to connect to our database with any existing Redis client or CLI. People can use this database without changing a single line of code in their application logic. If you want to see how databases actually become usable in the real world, give this a watch. 4 videos are out now: 1. Why Single-Threaded Redis Is Fast 2. Writing a TCP Echo Server 3. Wire Protocols 4. Implementing RESP ps: This video is part of my highly practical, in-depth, hands-on Redis Internals course, now being released on YouTube. Hope this makes you super curious about databases and engineering.
English
6
5
140
8.5K
Meghraj Deshmukh
Meghraj Deshmukh@itsdeshmukh·
@SumitM_X While all topics are valuable for learning, but I think should focus on learning or preparing for what aligns cracking good company's brand first. I know many people who are great at implementing,ICs but still at 20lpa with 7-8 years of exp
English
0
0
0
48
SumitM
SumitM@SumitM_X·
Want to be a backend architect ? Please learn : 1. Microservices Design Service decomposition, Bounded contexts, Resilience (Circuit Breaker, Bulkheads) 2. Distributed Systems Fundamentals CAP Theorem, Event sourcing, CQRS, Data consistency models (ACID vs. BASE) 3. High-Performance Data Management Database partitioning, Index optimization, NoSQL data modeling 4. Advanced API Design gRPC, GraphQL, API Gateways, Asynchronous APIs 5. Event-Driven Architecture Kafka, Message queues, Pub/Sub patterns, Saga pattern 6. Cloud-Native Patterns Container orchestration (Kubernetes), Serverless, Multi-cloud strategies 7. Observability Distributed tracing (OpenTelemetry), Centralized logging (ELK), Real-time monitoring 8. Infrastructure as Code Terraform, Helm, Configuration management best practices 9. Advanced Security Zero Trust, OAuth2, JWT, Data encryption in transit and at rest 10. Scaling Strategies Load balancing, Sharding, Horizontal vs. vertical scaling
English
8
135
864
130.6K
Priyansh Agarwal
Priyansh Agarwal@Priyansh_31Dec·
Apparently people aren’t able to finish DSA cause they keep starting with Array. The problem is that you treat DSA like Chemistry. You memorise and keep forgetting so you have to keep starting from scratch again and again. Even if you finish it, you will still have to do that entire process again. This is why the whole idea of revision is flawed.
English
34
51
1.2K
94.4K
Meghraj Deshmukh
Meghraj Deshmukh@itsdeshmukh·
@0xlelouch_ They are not judging on only an interview, it's your track record which speaks more than an interview. Interviews are important but for this kind of role previous track is more important
English
0
0
3
4.4K
Abhishek Singh
Abhishek Singh@0xlelouch_·
Razorpay Principal Engineer II (India) and people might be wondering “how much does it pay?” A senior of mine got an offer for principal engineer 1, So estimating from there and from what I’ve seen in the market (varies by team, location, stock refresh): 1. Base: ₹85L to ₹1.2Cr 2. Bonus: 10 to 20% (₹8L to ₹24L) 3. ESOPs: ₹40L to ₹1.5Cr vesting over 4 years (sometimes more if you’re a rare fit) 4. Typical all-in TC: ~₹1.3Cr to ₹2.5Cr Coming to the interview process, it's not “DSA grind only”. It’s heavy on judgment. 1. Recruiter screen (30 min) - scope, current TC, notice period, role fit 2. Tech screen (60 min) - 1 medium-hard problem - plus deep dive on how you think: tradeoffs, edge cases, complexity 3. 2 to 3 technical rounds (60 min each) 1. Coding: correctness + readability + tests 2. System design: payments-flavored problems (idempotency, retries, timeouts, reconciliation, queues) 3. Debugging and fundamentals: SQL, concurrency, incident style questions 4. Bar raiser round (sometimes) - “principal” expectations: leading across teams, simplifying architecture, killing bad ideas politely 5. Hiring manager + culture round - past impact, conflict stories, mentorship, how you drive execution If you’re aiming for this kind of level/role: go prepare 2 production stories with numbers (p95, error rate, cost), and be able to explain failure modes like you’ve been paged for them!
Arpit Bhayani@arpit_bhayani

Joined Razorpay as Principal Engineer II :) From being a long-time customer to now building parts of the system - it's a full circle. Fintech is a new territory for me - time to get under the hood of how money actually moves. New domain, same guarantees - availability, correctness, performance - just with real money on the line.

English
23
79
2.1K
335.8K
Praveen Neppalli
Praveen Neppalli@praveenTweets·
Agentic software engineering adoption is on fire at @Uber. 1,800 code changes per week are now written entirely by Uber's internal background coding agent, and 95% of our engineers now use AI every month across all the tools we track. This is a real reset moment for engineering; it's one of the most exciting times to lead. This shift requires builders to be curious and hands-on. I’m incredibly lucky to be surrounded by a team that’s doing exactly that. The best part is that the strongest adoption isn’t being pushed top down from leadership announcements; it’s coming from engineers who are quietly experimenting, quietly shipping, and quietly pushing things forward. I love spending time with those engineers because there’s no substitute for being close to the work. Over the last few months, we leaned in hard, and the results have been phenomenal. The bigger shift: going agentic. 84% of AI users are now working with agent-style workflows, not just tab completion. Claude Code usage nearly doubled in 2 months (32% → 63%), while IDE-based tools have largely plateaued. Engineers are moving from accepting suggestions to delegating tasks. Even within traditional IDEs, ~70% of committed code is now AI-generated. Background agents are writing code autonomously. Our internal background coding agent went from <1% of all code changes to 8% in just a few months. There is zero human authoring. Engineers review and approve, but the code is written entirely by AI agents. The role of the engineer is shifting - from writing every line to architecting systems and reviewing AI-generated code. More to come from the @UberEng team in the coming days.
English
179
168
1.8K
476.9K
SumitM
SumitM@SumitM_X·
@b_xyz_10342 There are multiple , what are your strong areas ?
English
2
0
1
520
SumitM
SumitM@SumitM_X·
In 2026 , you should - always be interview ready - have 2-4 income streams
English
12
23
436
19.1K
Ritesh Roushan
Ritesh Roushan@devXritesh·
Most likely queue or provider latency. The first OTP request gets delayed somewhere in the pipeline (SMS gateway, message queue, or carrier network). When the user clicks Resend, a second request is triggered and both messages get processed together once the delay clears. So it’s usually delivery lag in the messaging pipeline, not the OTP logic itself.
English
2
1
6
8K
Shreya
Shreya@Oblivious9021·
Send OTP → nothing arrives. Click Resend OTP → suddenly both OTPs arrive together. What’s the technical reason behind this?
English
255
486
10.2K
922.4K
Subham BRC
Subham BRC@SUBHAMBASUROYC1·
@LearnersBucket If software is indeed sunsetting, then why are you selling frontend courses?
English
2
0
15
2K
Prashant Yadav
Prashant Yadav@LearnersBucket·
The ERA of software is slowly sunsetting. Big techs will lean in every 12-18 months as AI evolves. The safe bet as of now is GCCs that are growing, but they do limited hiring. Keep observing for the next couple of months.
Aakash Gupta@aakashgupta

Software ate retail. Ate media. Ate finance. Ate transportation. Then AI ate $2 trillion of software in 30 days. Andreessen wrote “software is eating the world” in the Wall Street Journal in 2011. Most quoted sentence in venture capital history. He built a $43 billion fund around the thesis. Every pitch deck for a decade opened with that line. The food chain just flipped. Between January and February 2026, $2 trillion in software market cap evaporated. Not a recession. Not a rate hike. AI agents started doing the same work these companies charge $150/seat/month for. Salesforce down 30%. Workday 33%. Atlassian 35%, after enterprise seat count declined for the first time ever. Asana lost 59% in twelve months. DocuSign 52%. Jefferies coined it the “SaaSpocalypse.” The structural problem is brutal. SaaS runs on per-seat pricing. AI agents don’t need seats. When a company cuts headcount by 30%, it cancels 30% of its Salesforce licenses the same week. The product that ate every industry is watching its own customer base vanish. Goldman’s software basket trades at 22x forward earnings. That’s less than half the decade average. Price-to-sales compressed from 9x to 6x, levels last seen before the SaaS boom even started. Every software company spent 15 years eating someone else’s lunch. Now they’re on the menu. Naval said it in five words. The stock market said it in $2 trillion.

English
10
16
298
41.5K
Sunny R Gupta 🐰
Sunny R Gupta 🐰@sunnykgupta·
If there is interest, I can do a separate thread breaking down: How to prep specifically for some of these - a playbook for India‑based candidates (college + early career). 💬 Reply “PLAYBOOK” and I’ll put that together. Also, checkout: @TeamShiksha
English
40
3
46
5.5K
Sunny R Gupta 🐰
Sunny R Gupta 🐰@sunnykgupta·
🧵 If you’re in your early 20s, hungry, and actively looking for hustle culture (with real upside, not fake “startup vibes”), here’s a 2026 list of companies I’d aim for from India. Remember, this isn’t a “safe” or “complete” list ;) 🔖 Bookmark for later
English
35
98
547
53.2K
Marc
Marc@mrousavy·
nothing beats some high quality @margelo_com team time in the mountains 🇦🇹🏔️🏂
Marc tweet media
English
8
2
113
7.5K
Arpit Bhayani
Arpit Bhayani@arpit_bhayani·
You do not owe the world a side project. I know many great engineers with zero side projects, and many who became better because of theirs. So build only if you want to, not out of peer pressure.
English
49
53
1.5K
36.1K
Meghraj Deshmukh
Meghraj Deshmukh@itsdeshmukh·
@ujjwalscript But that's what business needs , refactoring is faster Building and throwing is cheaper then extending the perfect v1 Business people are pushing for this now Speed but expecting the quality as well
English
0
0
0
205
Ujjwal Chadha
Ujjwal Chadha@ujjwalscript·
Unpopular Opinion: We aren't building the future 10x faster with AI. We are just generating legacy code 10x faster. Everyone is currently bragging about developer velocity. "I built this entire backend in a weekend!" "AI wrote 80% of my codebase!" But here is the reality check we are ignoring: Code is a liability, not an asset. If an AI tool spits out 1,000 lines of functional boilerplate in five seconds, that is still 1,000 lines that a human being has to read, review, secure, and maintain when the dependencies inevitably break next year. We are treating code generation like a pure productivity win, but we are optimizing for the wrong metric. The bottleneck in software engineering was never how fast we could type. The bottleneck has always been comprehension, architecture, and maintenance. If we don't shift our focus from "generation speed" to "architectural sanity," the tech debt of the next five years is going to be an absolute, unmaintainable nightmare.
English
396
534
4.2K
637.7K
Neha Sharma
Neha Sharma@hellonehha·
please pivot from web (front-end) domain as fast as possible . It used to be a good money and easy entry role but now it is not and won’t remain. You will thank yourself in future.
English
16
1
155
9.9K
Neha Sharma
Neha Sharma@hellonehha·
@itsdeshmukh @Triggerman3311 Meghraj, don't say sorry to him/her/bot...he/she/bot is doing this for engagement..read his/her/bot twitter handler name - Tiggerman
English
1
0
1
16

Keşfet

@LearnersBucket @agent_wrapper @aoagents @composio @cneuralnetwork @championswimmer @arpit_bhayani @SumitM_X