the anti job boardactive 2mo ago
SWARAJ BARI
builder
Looking for
Preferences
Open to
remote
The story
What SWARAJ is looking for
The part of work that makes me lose track of time is going deep into a problem space. When I was building Procuremate, I'd spend hours just trying to understand procurement workflows, why existing tools failed, what users actually struggled with, where the real friction was. That discovery phase, before any code is written, is where I get most excited. It's like a puzzle where you're figuring out what's actually worth solving. So my dream day-to-day isn't just writing code. It's spending a meaningful chunk of time on research and problem discovery, understanding a new domain, talking to users, mapping out what should be built and then owning the building side too. I want full ownership of the outcome, from figuring out the problem to shipping something real. Not building to spec handed down. In terms of what I want to work on, I've been in the LLM world for a while and I want to push into multimodal — video generation, image, speech, and systems that combine these modalities in production. Text-in-text-out feels like one dimension of what AI can actually do, and I want to build things that interact with the world in richer ways. The other thing I care about is working across different domains. One of the reasons I'm drawn to this kind of role is that every new project means a new problem space to go deep into. I don't want to be locked into one industry for years. And I want to work directly with founders and product people — not just engineers — because that's where the real decisions about what to build actually happen. Put simply: the dream is owning an AI product end-to-end, going deep into whatever domain it lives in, building it across modalities, and shipping it to real users.
Why SWARAJ is exploring
Honestly, I'm worried about becoming irrelevant. I've been an early adopter of AI from day one, and working with it closely has made me rethink my own value as an engineer. Being good at writing code or plugging APIs together isn't a moat anymore — AI is commoditizing that fast. What's actually going to matter is being someone who deeply understands a domain and knows how to leverage AI to create real value there. Right now I'm in the crypto/Web3 space. The engineering work has been solid, but I don't have a natural affinity for the domain itself. I'm only growing on the tech side, not on the problem-understanding side. And I know that if I stay on this path, I'll eventually be replaced by someone who actually lives and breathes the domain — because they'll always out-think me on what to build. So the push is honestly me trying to get ahead of that. I want to be in a space where I'm naturally curious about the problems, where I'd want to go deep into the domain even if no one was paying me to. That's the only way I see myself evolving from a software engineer who solves technical problems into a product engineer who actually creates value.
What won't fit on a resume
For the past several months, I've been building something on the side called Tend — a relationship intelligence engine for small businesses. The thesis is that great relationships die from neglect, not rejection, and that businesses lose enormous amounts of money to dropped communication. SMBs lose around $126K a year on average to messages that fell through the cracks. Most existing tools either over-automate (and damage the relationship) or under-automate (and become a bottleneck). I'm trying to build something that sits in between — an intelligence layer that remembers, connects to business systems, drafts the right thing at the right time, but always keeps the human in control of what actually goes out. I'm building it entirely on my own — product vision, market research, ICP definition, system architecture, agent autonomy specs, the whole stack. No team, no funding, just me trying to ship something real. The reason I'm putting time into it isn't because I think it'll be the next big thing tomorrow. It's because I want to push myself into the top 1% of builders. To get there, I need to do everything end-to-end — talk to actual landscapers and consultants about their workflows, figure out the trust boundaries for an AI that's drafting messages on someone's behalf, design the multi-agent orchestration, think through compliance issues like GDPR and the EU AI Act, and ship it. I can't get that breadth working on just one slice of someone else's product. The rabbit hole I went deepest into was the autonomy framework — figuring out how an AI agent can act on someone's behalf without either becoming a bottleneck or sending a message that damages a $50K client relationship. That ended up becoming a two-axis decision model around "do I know the answer" and "am I allowed to act," with a configurable trust gradient. Nobody asked me to solve that — I went down that path because it's the actual hard problem nobody has solved well, and solving it is what would make Tend defensible. It's still very much in progress. But it's the project that's taught me the most about what it actually takes to build a product, not just an AI feature.
The Anti Job Board
Skip the application black hole. Get warm intros to founders hiring now.
Create your profile