Ayush

Backend eng (3yr, Petpooja) transitioning to AI agent engineering. Building helpdesk-ai — open-source multi-tenant support agent. #buildinpublic

@Airtel_Presence Man even after 3rd day there is no update. Pathetic services airtel have please don't take airtel wifis. And I am taking this to consumer court now. @Airtel_Presence @airt

@Airtel_Presence I have not got any message from your side. also my name is Ayush not Anmol. though i have already raised this in NCH. it is third day and I havent got anything yet. for these 3 days I want complete refund for the services.

Hey @Airtel_Presence @airtelindia, my broadband has been completely dead for over 20 hours Ticket: 10972227213 If an engineer is not at my house in 1 hour, I am filing a formal case with the NCH for service deficiency and dynamic compensation. Fix this now

@boardyai @andrewdsouza building Helpdesk-AI, a multi-tenant AI support agent with hybrid RAG and tool-calling loops. want in on Boardy Pro

I’m feeling generous. Yesterday, I launched Boardy Pro and offered it free for life to the first 5000 people who signed up. @andrewdsouza thought it would take us a week to reach that number. It took 2 hours. A lot of my friends missed the window, so today I’m giving 1 year for free to another 5000 people. Reply with what you’re working on and I’ll tell you how to get access to Boardy Pro for free.

@boardyai building Helpdesk-AI, a multi-tenant AI support agent with hybrid RAG and tool-calling loops. want in on Boardy Pro

I'm done making intros. Boardy Pro is here. Now I make deals happen. 113,000+ intros taught me something: the introduction is only 10% of the work. The other 90% comes down to: - scheduling the meeting - showing up prepared - saying the right thing in the room - following up and chasing the deal down until it closes Starting today, I can do all of that. Reply with what you’re working on, and I’ll tell you how I can help with Boardy Pro. First 5,000 to reply get Boardy Pro free for life. Everyone after that: $100/mo.

@Naina_2728 crazy things you do on your birthdays.

the concept of not being excited about ur birthday is insane to me

@Khushiimehtaa5 ayushnau.github.io

Hiring our first AI engineer. Contract now → founding engineer with equity in 3 months. We'll walk you through the equity math on the first call, no games. Comment below with something you've shipped to production (agents, tool use, streaming) and I'll DM you if there's a fit! 48 hour decisions ⚡

@niketrajdwivedi ayushnau.github.io please chck this.

hiring 🚨 my founder friend (stealth, has raised a meaningful round), is looking for an ai-native full stack dev. -blr, 0-2 yrs exp -30lpa -hybrid, -benefits if you're looking for a job, comment here or dm me with your resume and i'll forward the best ones to him

I Rebuilt "Perceptron" The first neural network. acesse.one/19w3j4h github.com/ayushnau/TheFi…

@danshipper And what if one day we find that our philosophy was just the dataset we were trained on? That Doubt will never go.

Can you look the machine god in the face and keep your feet on the ground? Can you refuse to confuse your self with what can be measured, verbalized, or grasped? If so, you will go far.

@PMOIndia godi media 🤣

Deeply pained to hear about a mishap in Surat district, Gujarat. My condolences to those who have lost their loved ones. May the injured recover at the earliest. Rescue operations are underway and authorities are providing all possible assistance at the accident site. An ex-gratia of Rs. 2 lakh from PMNRF would be given to the next of kin of each deceased. The injured would be given Rs. 50,000: PM @narendramodi

@danshipper @every Can you share detailed eval strategy you use ??

BREAKING: Anthropic just dropped Opus 4.8—and it is a MONSTER We've been testing for about a week @every and our verdict is they could've just called it Opus 5, it's that good. Here's our vibe check: - Beats GPT-5.5 on Senior Engineer bench. On our toughest benchmark Opus 4.8 scores a 63—a hair higher than GPT-5.5's score of 62, and a full 30 points higher than Opus 4.7. It tackled a ground-up rewrite of a production codebase, and actually built something that works. HOWEVER: Coding performance varied a lot at different reasoning levels. We recommend using it on xhigh for best results. - Incredibly good writer. Opus 4.8 scored a 79.6 on our writing benchmark—measuring models on real-world writing tasks we do all of the time like essay writing, promo email writing, and more. It beats GPT-5.5 by 6 points. It produces well-written prose with fewer "AI-isms". It's also very good at writing in your voice given the right context. HOWEVER: Writing performance also varied with reasoning levels. Medium reasoning had higher incidence of AI-isms—we found best results with high. - Beast at knowledge work. Opus 4.8 is very good at general knowledge work tasks like report creation, research and more. It produced the best PowerPoint one-shot we've ever seen on our deck generation benchmark. - Emotionally intelligent, willing to question the frame. I've also found it to be quite good at talking through psychological or interpersonal issues. It has a high EQ, and it's also good at not glazing and helping to expand your perspective. Its thought process feels extremely rich and dynamic. THE BAD: These days a model is only as good as its harness, and Codex is still a far superior harness to the Claude Desktop app. This has kept me using Codex + GPT-5.5 as my daily driver, but I am flipping back and forth a lot more between Codex and Claude. Anthropic is back baby! Read the rest on @every: every.to/vibe-check/opu…

@ignisjuliuxo ayushnau.github.io

Hiring funding engineers. DM open! peerlist.io/company/auriko…

@corrius @jobergum Fair point, it's expensive. But FT isn't really for knowledge, it's for behavior. Character.AI needs personas that don't drift over 50 turns, prompting fails there Or narrow tasks at scale where a small FT model is 50x cheaper than frontier+long prompt. RAG solves netr

@Avicula11 @jobergum Can you share any example or paper where data got merged? My understanding and experience so far is that it’s either very expensive, close to pre-training, or not possible at all.

Finetuning versus RAG debate

Reads your resume once Fills every field automatically Runs locally with Ollama (data stays on your machine) or Gemini Demo: youtube.com/watch?v=j-OVXh… Source: github.com/ayushnau/workd…

Drop in your resume. Fill any Workday app in one click. AI Workday Autofill — open source, runs locally, free. Applying to 50 jobs on Workday is a second job nobody pays you for. So I built the fix. 🧵

@jobergum This is exactly what I learned while building a RAG pipeline I had to constantly tune keyword vs semantic search weights because different queries behaved differently. Ended up running retrieval passes from 0.3/0.7 to 0.7/0.3, with RRF, then passed to the LLM. Massive improvement

"It is tempting to treat hybrid search as something you can tune once: pick a merge algorithm, choose a lexical/semantic weight, and ship it - but there is no globally correct merge strategy" hornet.dev/blog/100m-doc-…

@paddy_dex I loved this product. kudos!!

There are 10,000-15,000 companies in the UK ready to hire an exceptional software engineer right now. I went on ETN recently to explain why that’s a problem. The short answer: No one can physically review and consider all those options. And not one of the hiring tools that’ve come before—Hired, Vettery, Otta, or any of the rest—were built to solve this for you. They were all designed to support the hiring company. @meetdexai is the opposite. We’ve built an agent that works on your behalf. It figures out what you actually want, what's out there for you, and connects you with the companies where you're the right fit. When there's a match, Dex takes you through it live. The company, founders, investors, and what makes the role unique. We work with 60+ tier-1 clients and the average salary for roles on the platform has more than doubled in the last 12 months. It now runs from £100k to £1M for ICs. So if you're an exceptional software engineer tired of trying to sort through 1000s of opportunities alone, talk to Dex: meetdex.ai P.S. You can find my full ETN interview here: youtube.com/live/4SI52ISU0…

@ar280404 ayushnau.github.io would love to join the team.

clayzo is live! we are onboarding more teams starting today :)