the meh - vijay
nothing special here. nothing that you can remember me for. (unless you know me personally me) or maybe, i’m just good at hiding in plain sight. i build, i break, i learn, i move on. sometimes, i try create things that change the way people work. sometimes, i scrap things because they aren’t fun anymore. but does any of it matter? not really. i am just another mind trying to make sense of the madness that is existence, leaving behind trails of automation, ai models, and half-written thoughts that no one will ever read.
the path i’ve carved
miserable? maybe. directionless? perhaps. but stagnant? never. i’ve never been one to settle, to stay in a comfort zone, to follow a path paved by others. from the first time i wrote a script to make my life easier to deploying full-scale ai systems, every step was a rebellion against convention. i never aimed to be a part of the system; i wanted to break it, rebuild it, and reshape it in ways that made sense to me.
a little walkthrough of professional path i’ve walked till now
- [feb 25 - march 25] - themadcollective previously called plaigroundhq
- iterative surveys - realtime voice + text based interactive and iterative surveys
- upscaler - automations, like bulk image generation and upscaling
- seamless design pattern generators for textile industries,
- [march 24 - jan 25] - pragetx
mostly did ml and ai did some backend (django, fastapi, flask), frontend(nextjs, reactjs), and little bit of other software dev stuff (cpp, flutter, node) etc.
projects i worked on there
- spanish tts - trained a spanish real like tts + voice cloning
- the people’s economist - created finacial tools lots of web scraping and rag based chatbot and for analysis/advisory of your funds/portfolios.
- voagents - end to end ai voice assistants, integrating vad, stt, llm, tts, tool calling.
- bejoice - container shipping - created a library for container packing with python + c++, threejs
- dynamic-csv-to-sql - converting dynamic input information to incrementally upgrading sql database using sqlcoder and llms + classification task.
- llm eval - deploying finetuned llm using vllm to evaluate the accuracy + performance on translation and classifying the text that doesn’t makes sense or may change in meaning if translated.
- website-chatbot - a chatbot with web navigation assistant.
- real estate recommendation - scraping dubai properties, analytics (kibana, elasticsearch), search recommendation (clustering) etc.
- astrosphere - backend + deployment of astrotalks clone, realtime convos, chats, ws
- infinititalks - similar live chat/call app for psychologists/dehypnotherpists
- pragetx website - ssr in nextjs and seo + deployments and ci/cd
- lighthouseagencyinc - ssg in nextjs deployment on cloudfront and s3
- purna paper products - ppp - backend for a distribution app for sellers, delivery partners & shopkeepers
- gst toolkit - a simple window app for easy financial calculations for specific usecase.
- [nov 23 - jun 24] - quantumreach - did webscraping for lead generation started a startup on autoblogging system (recieved vc grant actually), but a wrapper wouldn’t last long
- [aug 23 - oct 23] - trakky -
- web dev, django backend, react frontend, ec2 deployment for a salon aggregator
- [dec 22 - may 23] - rp-hplc - ljinstitute of pharmacy - rp-hplc - ml model development for a research project for drug mobile phase prediction for reverse phase high performace liquid chromatography
i have walked through codebases that felt like labyrinths, boxed with infrastructure that refused to cooperate, and battled the constraints of time and sanity while pushing projects from ideation to execution. company classifier, autoblogging systems, real-time call assistants—none of these were planned. they were spontaneous sparks, fueled by curiosity and an undying need to build something that simply did not exist before.
maybe i haven’t carved a traditional path. maybe i never will. but in every broken function, every late-night commit, every scrapped prototype, there is a story. a story of trial, error, and an unrelenting desire to see what happens when you refuse to do things the normal way.
meet in the middle
i hate people, but have faith in individuals. people are ugly. but you might not be. i want to meet the cool people, the magicians, the makers, the creators of the mirai(the future), and the carrier of the kako(the past). i might not be able to find you, but if you do find me, reach me, at least say hi, or hey or whatever, cuz i and maybe you might regret a possible life or at least a short fun story that we might have lived together.
the breadcrumbs i’ve left behind
if for some reason you need to find me—maybe to talk tech, share ideas, or just yell at me for breaking something—here’s where i exist in the digital void. no guarantees i’ll respond, but who knows? maybe the stars will align.
scribbles in the void
: wandering in the matrix (github)
: my corporate facade (linkedin)
: summon me (call)
📞 : +919898864226
send words into the abyss (email)
: my corner of the web (website)
🌐 : huggingspace
🌐 : h3yfr13nd.live
🌐 : h3y.xyz It’s “hey whoever”
broken models & experiments (huggingface)
: numbers & data playground (kaggle)
: other places i forget i have accounts on
🌀 : uhh. i forgot.
the way i am
i don’t always know why i do things. sometimes, it’s a burst of excitement, an itch to create, or sheer boredom knocking at my door. but when i look back, i see patterns—patterns of reckless curiosity, whimsical problem-solving, and a stubborn refusal to do things the way everyone else does.
people might call it erratic. i call it freedom.
i build things, break things, fix them differently, then break them again just to see if i can make them better. i never wanted to follow a well-trodden path because, honestly, where’s the fun in that? so, i carve my own—one weird, unpredictable step at a time. whether it’s designing ai systems, automating workflows, or spinning up entire platforms overnight, i thrive in the madness. the lines between work, play, and obsession blur so much that i often forget which is which.
i’m not a methodical planner. i don’t meticulously map out every detail before starting. i dive in, experiment, and let the process teach me. some call it reckless, and maybe it is. but i’ve learned that real innovation doesn’t come from careful steps—it comes from bold leaps. the world doesn’t reward caution; it rewards those who dare.
every project, every model, every line of code i write isn’t just about getting things done—it’s about proving to myself that there are no limits. that i can push the boundaries of what’s possible. that the only thing standing between an idea and its reality is the will to build it.
sometimes, i wonder if i should slow down, strategize more, plan better. but then i realize—i don’t want to. i love this. the spontaneity, the trial and error, the late nights chasing a thought that won’t let go. it’s not about efficiency. it’s about creation. and in that chaos, i find clarity.
maybe no one else will read this. maybe this is just me talking to myself, reminding myself why i do what i do. but if i ever forget—if i ever find myself doubting, hesitating, or getting caught up in conventionality—this is my reminder:
i don’t follow. i create. and i wouldn’t have it any other way.