Website:
Job details:
Backend Engineer: Part-time
Early-stage startup · NDA required
We're building Maaya AI: a real-time voice AI operating layer for busy professionals. Think of it as a persistent judgment layer that handles communication overhead so founders can stay in execution mode. We're in beta. We're moving fast. The window to build this before Big Tech closes it is real.
We need a backend engineer who treats constraints as design challenges, not blockers. This isn't a ticket-closing role. You'll be making architecture decisions.
What you'll work on:
Voice pipeline on edge: Deepgram STT + TTS, GPT-4.1 primary, Gemini 2.5 Pro fallback
Latency optimization every 100ms matters in voice UX
Persistent memory architecture, context threading, session state
API integrations: calendar, email, communication tools
Reliability and failure handling for real-time voice flows
You're the right fit if:
Strong in Node.js or Python backends, REST APIs, WebSockets
Experience with LLM API integrations (OpenAI, Gemini, or similar)
Comfortable with ambiguity: specs will change, priorities will shift
You've shipped production code, not just personal projects
You push back when something is architecturally wrong, not just when asked
What we offer:
Part-time, flexible hours: we care about output, not availability windows
Equity conversation if you're serious about the long game
Direct access to the founder: no PM layer, no ticket queues
You'll have actual influence over how this product is built
One thing to know upfront:
An NDA is required before any detailed technical discussion. We're building in a competitive space and we protect what we're building. If that's a dealbreaker, we're not the right fit for each other.
If this sounds like your kind of problem: DM me or drop a note at jpatel@duchastech.com with a quick intro and something you've built that you're actually proud of.
Click on Apply to know more.