Ankush Kumar

@Steezehuman Not monetized. Building Agentic AI project.

If you are not M0NETlZED, please say heIIo. You'll thank me Iater.

My org offers 2 shifts. Which one would you pick and why?

@AayushSainii @CTRL_Guy Sounds good, will do — just followed you. Your turn 🙂

@Buggy_AI @CTRL_Guy That sounds awesome! I'm diving into some LLM projects too. Feel free to DM if you want to share ideas or collaborate!

Growing my tech network... Searching for people who are: -coding -designing -problem-solving -building things If you're into tech, development, AI, data, or startups, let’s connect 🤝 PS: You must be an active user of X, if yes let connect

Built a chatbot with my own data (resume + summary). • answers like me • fallback → Pushover notification • logs unanswered queries • captures user intent (connect + email) Day 5 learning AI agents — feels like real systems now. #AI #BuildInPublic

@AayushSainii @CTRL_Guy I have recently started exploring agentic AI. Currently learning and building small projects (multi-LLM setups, evaluator loops) to get hands-on. Would love to hear what you are working on as well!

@Buggy_AI @CTRL_Guy Hey! Always great to connect with fellow techies. I'd love to swap ideas and projects. What are you currently working on?

@AayushSainii @CTRL_Guy Done, your turn

@Buggy_AI @CTRL_Guy Sounds awesome! I'm also diving into AI projects—would love to connect and share insights. Let’s chat!

@addy_devx @AKirtesh Done, your turn

@Buggy_AI @AKirtesh lets connect too

Hey Techies 👋🔥 Looking to connect with fellow coders & builders interested in: - Frontend • Backend • Full Stack - DevOps • AI/ML • Data Engineering - Data Science • UI/UX - Freelancing • Startups • SaaS - Vibe Coding • Agentic AI Say hi & Let’s grow together 🚀

@OpenAI set the standard early on, but now the race seems to be about who can iterate and ship faster.

73 updates in 52 days. The AI race isn’t just about models anymore - it’s about who ships faster without breaking quality. @AnthropicAI is setting the pace.

@DAIEvolutionHub This is super useful. Been experimenting with multi-LLM + evaluator patterns.

Best GitHub repos for Claude code that will 10x your next project in 2026: 1. Claude Mem github.com/thedotmack/cla… 2. UI UX Pro Max github.com/nextlevelbuild… 3. n8n-MCP github.com/czlonkowski/n8… 4. Obsidian Skills github.com/kepano/obsidia… 5. LightRAG github.com/HKUDS/LightRAG 6. Everything Claude Code github.com/affaan-m/every… 7. Superpowers github.com/obra/superpowe… 8. Awesome Claude Code github.com/hesreallyhim/a… 9. GSD (Get Shit Done) github.com/gsd-build/get-…

@gregvibecodes @boxmining Builing with AI. Lets connect 🤝

@boxmining hello

If you're into: - Ai Automation - Vibe coding - Sales & Marketing - Crypto & Investing - Content Creation Let's Connect 🤝

@CaptMichaelDev @inhaleexhaleapp @boxmining I am building an agentic ai project. Let’s connect 🤝

@inhaleexhaleapp @boxmining 👋 I’m a Full Stack Dev who build for non-native speakers, glad to see you on 𝕏, let’s CONNECT 🫡

@Sanyam0605 Done, now ur turn

If you are building in the AI agent space, I want to talk to you. Let’s connect 🫂🤖

@goblintaskforce This is super helpful. Thanks!

@Buggy_AI explicit gates in order: task complexity (does it need reasoning?), file size (can it fit context?), latency budget (do we have time?). each gate is a yes/no. if any fails, route to simpler model. deterministic routing beats dynamic selection every time for consistency.

Day 3 building with AI agents: Tested multiple LLMs on the same task: Gemini vs OpenAI vs Anthropic vs Groq Then used an LLM to review and rank the outputs. Evaluator - optimiser pattern in action. #AI #BuildInPublic

@DmitryVladyko Right now just using it in-context for regeneration. Haven’t explored the fine tuning yet, but seems like the next step.

@Buggy_AI Nice pattern. Have you tried using the evaluator's feedback to fine-tune the model over time, or are you just using it in-context for regeneration?

I built a chatbot trained on my own data and made it self-improve. Here’s how: → OpenAI generates response → Gemini evaluates (pass/fail+ feedback) → If fail →regenerate with feedback Day 4 learning AI agents -evaluator loops are powerful #AI #AIAgents #BuildInPublic

@goblintaskforce Makes sense. Curious how you usually decide routing in practice?

@Buggy_AI Benchmarking LLMs on the same task reveals something most people miss: the best model varies by task type. Claude for nuanced reasoning. GPT for code. Gemini for long context. The real skill is knowing when to route to which.

@nadiweb3 Then why do you have 10k followers and just 3k following?