Built by engineers who saw a better way

The story of how two founders at a hackathon decided to fix enterprise knowledge management

It started at PennApps

We met at a hackathon while we were both at UPenn. Within the first conversation, we realized we were both itching to build something real—not just a side project, but an actual company that solves a problem too big to ignore.

Our skills just clicked. Nirav had already gone through the grind of running a Delaware C-Corp and knew the B2B SaaS world inside out. Neel was the backend architect who got excited about hard technical problems that most people would avoid.

At that hackathon, we shipped our first thing together. Nirav handled the full-stack work and product decisions. Neel built out the backend and infrastructure. It just worked. No friction, no awkward handoffs—we naturally complemented each other's strengths.

By the end of the weekend, we knew this wasn't just a hackathon partnership. This was the beginning of something bigger.

Every company has the same broken system

Before building, we talked to managers across different industries—tech startups, financial firms, consulting companies, universities. We asked them what's keeping their teams slow and why they're not adopting modern AI tools despite knowing they need them.

The same problems came up over and over:

Information is scattered everywhere

Google Drive, Slack, email, Notion, Confluence, SharePoint—knowledge lives in a dozen disconnected places and nobody can find anything when they actually need it.

Search doesn't work

Traditional search matches keywords, not meaning. You search for "pricing strategy" and get file names with those words, not the actual strategic documents you need.

AI tools can't access company data

ChatGPT is powerful but has zero context about your business. Enterprise AI products either can't connect to your systems or require months of implementation with consultants.

Security blocks everything useful

Companies are terrified of putting sensitive data into third-party AI tools, so they just… don't. The tools that could help the most sit unused because nobody trusts them with real information.

Knowledge walks out the door

When Sarah from Finance quits, her entire knowledge base disappears. When the engineering lead leaves, six months of context evaporates. Companies have no institutional memory.

We realized this isn't a productivity problem—it's an infrastructure problem. And nobody was building the complete solution.

Deep technical builders who understand the problem firsthand

Neel Manro - Co-CEO

Backend architect and infrastructure specialist. The person who gets excited about building systems that scale to millions of documents without breaking.

Neel owns:

• Backend architecture and APIs
• Database design and optimization
• Our proprietary retrieval engine
• Infrastructure and deployment
• System performance at scale

Before Ziqara, Neel experienced these problems watching teams waste hours searching for information that should be instantly accessible, seeing companies struggle to adopt AI because it couldn't connect to their actual data.

Nirav Rohra - Co-CEO

Full-stack developer and product strategist. Has run a Delaware C-Corp before and knows B2B SaaS inside out.

Nirav owns:

• Full-stack development
• Product design and UX
• Enterprise integrations (Google, Microsoft, Slack)
• Go-to-market strategy
• Customer experience

Before Ziqara, Nirav watched knowledge walk out the door when people left companies, saw teams re-create work that already existed somewhere in the system, and knew there had to be a better way.

Together:

We've written every line of code. Designed every feature. Made every product decision. Nothing's been outsourced or built by anyone else.

We're technical founders who can build anything. Now we're focused on building the one thing companies need most: a system that actually remembers and understands their business.

Two months from concept to enterprise-ready

We officially started working together two months ago, but we were both already solving similar problems independently before we met. When we connected at PennApps, we realized our visions aligned perfectly, so we merged what we'd built. About 40% of the work was already done between us.

Over the past two months, we've been working full-time—8-9 hours a day minimum, most days even longer. We're in different cities (Surrey, BC and Dallas, TX), but we're on calls constantly, doing long work sessions together and moving fast.

In those two months, we took the product from 40% to fully polished and enterprise-ready.

What we built:

Near-zero data loss ingestion engine

Parallel processing architecture with 128 concurrent workers. Handles PDFs, spreadsheets, images, audio, scanned documents—everything. Intelligent OCR, streaming ingestion, fault-tolerant recovery. Your files don't just get stored—they get understood.

The most accurate retrieval system on the market

Proprietary hybrid search combining six different methodologies: dense vector search, sparse retrieval, semantic understanding, cross-document reasoning, metadata signals, and user context. Typical query latency: 1-3 seconds across millions of documents.

Enterprise-grade security

Three-tier permission system with database-level isolation. Team files stay locked to that team. When you delete something, it's gone forever—no hidden backups.

Native meeting intelligence

Built-in transcription, summarization, and permission-aware storage. Start a session, select access level, and we handle the rest.

Real integrations

Google Drive, Slack, Microsoft Teams, OneDrive, Jira, Confluence. Not just connecting—actually understanding and traversing your entire knowledge ecosystem.

We're both all-in on this.

Beyond enterprise—reimagining how organizations learn and operate

We're starting with enterprise knowledge management because that's where the pain is most acute and the market is ready. But this is just the beginning.

The same architecture that powers Ziqara for companies will transform education. Universities are stuck with Canvas, Moodle, and Blackboard—outdated learning management systems that are glorified file storage with discussion boards.

We're going to replace them with an AI-native learning platform where students and professors interact with a shared institutional knowledge brain. Instead of static syllabi and PDF dumps, classes become intelligent, adaptive conversations. Knowledge doesn't just get stored—it gets connected, contextualized, and made alive.

Our long-term vision:

Every organization—companies, universities, hospitals, governments—needs a system that captures, preserves, and makes accessible their institutional knowledge. Not just search. Not just storage. Not just AI chat. The complete cognitive infrastructure.

We're building the foundation for how organizations will operate in the next decade. The shift from storing information to actually using it. From scattered tools to unified intelligence. From knowledge that resets when people leave to institutional memory that compounds over time.

This is infrastructure, and infrastructure shapes everything built on top of it.

We know how to build. We need to learn how to win.

We're building in the fastest-moving space in tech right now, and every week we wait is a week our competitors get ahead. The AI race isn't won by having the best technology—it's won by who executes fastest and gets to market first.

We know how to build. We built a production-ready enterprise RAG system that actually works. We deployed to real infrastructure. We optimized our database for scale.

But building is the easy part.

What we don't know is how to go from zero to signing 100 companies in 12 months. Y Combinator has the playbook for that—the go-to-market strategy, the intro to the first 10 customers, the pressure to move faster than feels comfortable.

We're technical founders who can build anything. We need YC to teach us how to win the market before someone else does.

Building for the long term

We're not here to build a feature and get acquired next year. We're here to build infrastructure that matters—the kind of company that's still growing in a decade.

What that means:

We stay technical

We're not handing off engineering to hire more salespeople. We write code. We make product decisions. We understand our stack at the deepest level because that's how you build something that lasts.

We solve hard problems

Near-zero data loss ingestion. Six-engine hybrid retrieval. Team-level security isolation. Meeting intelligence. These aren't easy problems. That's why we're solving them.

We listen to customers

We have three companies testing Ziqara right now. We're learning fast and iterating based on real usage. When they tell us what they need, we build it.

We move fast

Two months from 40% to enterprise-ready. That's our pace. We're not slowing down.

We think big

Enterprise knowledge management is a billion-dollar opportunity. Education is another. This technology applies to every organization with more than 50 people. The TAM is enormous and we're just getting started.