Aamir Orbit
WritingSenior Product EngineerAI × Web3 Security
I build products, startups, and AI systems — from the first idea to the people actually using them.
Seven years shipping across AI, fintech, health, and web3: an AI receptionist for clinics, a therapy marketplace, a prediction market, a smart contract firewall. I do the architecture, the build, the deploy, and the part afterwards where you find out what you got wrong.
Projects
An AI receptionist for med spas. It answers calls, texts, web forms, and Instagram DMs inside sixty seconds and books straight into Boulevard, Acuity, or Calendly — then reports weekly on the revenue it recovered.
A prediction market for retail traders across Asia. YES/NO positions with instant odds and defined payouts, priced by a constant product market maker and settled on-chain. MPC vault wallets handle USDT, in seven languages.
A job board that knows who you are. Roles at newly funded AI and crypto startups land within hours of the announcement, often before LinkedIn, scored against your actual resume and ranked by fit.
A community-driven identity and reputation platform for the Base ecosystem. Earn Visa Points for verified contributions and work toward full Citizenship — so builders find each other on what they have shipped, not on follower counts.
Online therapy for Pakistan, matched rather than listed. A five-minute intake proposes two to five licensed psychologists by concern, preference, and language; weekly video sessions with follow-up messaging in between, end-to-end encrypted.
A couples app for people who are not in the same place. Moods sync instantly to your partner, letters and shared games sit next to a real-time collaborative canvas, and lock screen widgets keep it visible without opening anything.
Smart contract security infrastructure, built as three systems. Monitor watches live transaction flows for anomalies, Interceptor analyses them and executes preventive action, and Guard is a modular firewall enforcing geo-verification, threshold approvals, and time-locked execution before a transaction can sign — programmatically, through Gnosis Safe.
RAG Knowledge Platform
A central knowledge hub serving several technical initiatives. Queries are classified by intent, retrieved by vector search, re-ranked so recent material wins, then streamed back token by token.
Multi-provider agent runtime
One agent surface over GPT, Claude, and Amazon Bedrock, with dynamic provider routing and failover handling — so a provider outage does not become a product outage.
Multi-tenant AI SaaS & SDK
The platform underneath the above, plus the 30+ endpoint SDK on top of it: JWT auth, role-based access, per-tenant rate limiting, API-key lifecycle management, and multi-environment support.
Work
Product Engineer, Stealth Startup
Own the full lifecycle: architecture, build, deploy, adoption. Founded a community-driven web3 identity and reputation platform, and translated zk-EVM, zk-SNARK, private app channel, and threshold-signature research into shipped products and developer experience.
Full Stack AI Engineer, Ethlas & FailSafe
Led backend architecture for FailSafe. Built the agent runtime, the contract-intelligence system, and the multi-tenant platform they run on.
Software Engineer, DEVSiNC
Built and led a backend team. System architecture, scalable APIs, secure data and deployment pipelines.
Technical Product Engineer, The Social Cubes
Scalable web applications, API integrations, backend performance.
Stack
Writing
You Cannot Prompt Your Way to Safety
Web3 learned that you cannot secure an irreversible action with a warning message, because there is no undo. AI agents are acquiring the same property, and the industry is rediscovering the same four controls under new names.
Your First Engineering Hire Should Think Like a Forward Deployed Engineer
Forward deployed engineer job postings grew eightfold in a year. The role gets framed as an enterprise problem, but the mindset matters more at five people than at five hundred.
Failover That Loses State Isn't Failover
Two LLM providers at 99.53% each give you 99.99% combined. That arithmetic is real and it is also the easy half of the problem — because the conversation does not survive the switch.
What Breaks When Your AI Product Meets Its First Enterprise Customer
It is not the model. It is that a request stopped being a unit of cost, one customer can now exhaust another customer capacity, and someone is going to send you a security questionnaire.
Your RAG System Thinks It's Always Today
Embedding search has no opinion about time. It will hand you a policy that was superseded eighteen months ago with exactly the same confidence as the one that replaced it — and your users will not be able to tell.