Backend Engineering, Databases & System Design made simple.

terminal
Joined December 2025
D3 Bytes retweeted
Asked at Microsoft, Databricks, Glean & Palantir 🤔 You’re deploying a 70B-parameter LLM in production. How much GPU memory do you actually need to serve it?
51
20
1
132
11,958
D3 Bytes retweeted
Edgex
@SahilExecOct 23
bro asked claude for an office and got a whole GTA side quest 💀
I asked Claude to build an office for the @tavus PALs... I was not expecting it to build THAT kind of office
26
5
38
1,048
D3 Bytes retweeted
Interviewer: If I fine-tune an LLM on my company's data, do I still need RAG?
130
25
4
457
74,299
D3 Bytes retweeted
Interviewer: If I increase an LLM's context window from 32K to 1M tokens, does that mean it can reliably remember everything in that context?
113
23
2
341
86,608
D3 Bytes retweeted
Edgex
@SahilExecOct 16
3D generation is entering its Agent-Ready era. Hyper3D is leading the shift. I created a production-ready 3D camera using Hyper3D Rodin Agentic Mode. Instead of manually deciding the workflow, Agentic Mode analyzed the reference, selected the best generation approach, and handled the heavy lifting automatically. Workflow: 🔹 Upload reference 🔹 Agentic Mode chooses the optimal modeling path 🔹 Generate the 3D camera 🔹 Refine and iterate with natural language * Parameter Controls -- adjust dimensions and add custom controls for specific parts. * Animation Presets -- describe a movement or gesture and generate an animation with natural language. * Enhanced Materials -- modify the material of specific parts while keeping the rest of the model intact. 🔹 Export production-ready results In this demo, I show the complete process from a camera reference to a fully realized 3D model and the final refined result. Try Agentic Mode on hyper3d.ai #Hyper3D #DeemosTech #Rodin #AI3D #agentready #3DModeling #Agent #agenticmode
28
6
30
1,211
D3 Bytes retweeted
Agentic AI interview question: The agent claims it successfully updated a customer’s address, but the database was never changed. How would you design the system so the agent can’t declare success without evidence?
64
18
2
125
18,937
D3 Bytes retweeted
Edgex
@SahilExecSep 21
3D generation is moving from “generate an asset” → “build and refine the asset.” @DeemosTech’s Hyper3D Agentic Mode can generate editable 3D models, adjust dimensions and proportions through natural language, add animation, and even combine multiple reference views into one coherent model. That’s a much more interesting workflow than one-shot generation - you create, talk to the model, refine, and keep building. If you work with 3D, CAD, games or product design, this is worth checking out.
🚀 Introducing #HYPER3D Agentic Mode. AGI optimizes your input and chooses how to model it. Generate editable N-GONS #3D model for #CAD, building & industrial product. 📏Adjust dimensions. 🏃‍♀️Add animation. Towards Production-Ready, to Agent-Ready. This is Vibe Modeling👇
21
7
25
1,296
D3 Bytes retweeted
Just wrapped up my backend project an e-commerce system, but not the basic CRUD kind. Node.js + Express, JWT auth with refresh rotation, Redis caching + locks, Razorpay webhooks with HMAC verification, atomic stock reservation. Would genuinely appreciate feedback from other backend devs what's missing, what you'd do differently, what breaks at scale. Also open to suggestions what's a good next backend project to build? repo: github.com/devdeepdhariwal/e… (don't tell me "JS is bad for backend" tell me what a beginner should actually start with instead )
🤖 Made with AI
1
3
7
164
D3 Bytes retweeted
3
97
12
1,908
239,631
D3 Bytes retweeted
3
18
3
267
29,280
D3 Bytes retweeted
38
475
16
6,994
226,778
D3 Bytes retweeted
100 followers in 2 days 🎉🥹 Thank you everyone for being part of this journey. Looking forward to connecting with more people in tech. ⚡ Backend 🎨 Frontend 💻 Full Stack ☁️ DevOps 📊 Data Science 🎯 UI/UX 💼 SaaS 🤖 AI & Vibe Coding 🟨 JavaScript 🧩 DSA
7
2
9
345
D3 Bytes retweeted
238
7,977
329
49,067
1,822,796
D3 Bytes retweeted
5
97
1
1,201
32,212
D3 Bytes retweeted
3
3
9
1,108
D3 Bytes retweeted
51
478
14
12,375
484,705
D3 Bytes retweeted
76
3
1,468
76,269