Deepanshu
97 posts

Deepanshu
@DeepanshuCode9
Mern ll Java Full Stack Developer @sheryians https://t.co/eWe1lkMOa6
india Haryana Katılım Eylül 2021
51 Takip Edilen13 Takipçiler

Built a Perplexity-style AI chatbot
• Real-time AI responses
• Clean chat UI
• Markdown rendering
• Full-stack architecture
Still improving the UX, streaming & response flow
#AI #LLM #BuildInPublic
English

Built SnapCart - a full-stack MERN e-commerce app
• Users can browse & buy products
• Sellers can create accounts & sell
• Google OAuth + Razorpay payments
• Clean UI + real backend logic
Live: snapcart-7kgp.onrender.com
#BuildInPublic #MERN
English

@pavelhegler That’s the next step
Focusing on onboarding first sellers and validating the flow in real use.
English

Built product creation API for my e-commerce project 🚀
Implemented:
• Protected seller-only route
• Image upload with ImageKit
• Product validation
• MongoDB product creation
Seeing that 201 Created hit different 😄
#BuildInPublic #MERN #BackendDev

English

Learned today why planning before coding matters.
Mapped an e-commerce seller flow into models, APIs, middleware, validation & protected routes.
Architecture matters as much as implementation.
#BuildInPublic #WebDev

English

Built auth for my new e-commerce project
• Login/Register UI
• Backend Auth
• Google OAuth
• Frontend Connected
How can I improve this UI/UX?🤔
#BuildInPublic #MERN #UIDesign


English

🐛 Can you spot the bug?
Building a Banking Ledger in Node.js something in this code could break in production
👀 Look closely at the controllers...
#SpotTheBug #NodeJS #JavaScript #CodingChallenge

English

Building a transaction system with idempotency
→ Prevent duplicate transactions
→ Validate sender & receiver
💵 Ensure safe money transfer
Learning how backend systems handle reliability.
#NodeJS #Backend #BuildInPublic

English

@AgiletTask_ That makes a lot of sense, thanks
Structuring the input before sending it to the LLM seems like a big unlock.
I’ll try implementing section-based parsing and see how it impacts output quality.
English

@DeepanshuCode9 Section-based parsing is the right call — treating
the resume as structured zones rather than raw text
makes the LLM's job much easier.
We do something similar in AgileTask for sprint
generation — structure the input before the LLM
sees it and quality jumps significantly.
English

Building an AI Resume Analyzer 🤖
→ PDF → LLM → Interview insights
→ Questions + skill gaps
Improving prompt structure for better results.
#AI #NodeJS #BuildInPublic

English



