Senior Software Engineer - AI Hardware Backend

Nuha; New York (NY); On-site; USD 120,000 - 160,000; Full time

Job summary

Nuha is seeking a Senior Software Engineer to develop the backend infrastructure driving our AI hardware. In this role, you'll create FastAPI services for managing interactions between our embedded devices and intelligent services.

Applicants should possess strong systems thinking, experience with real-time data pipelines, and a collaborative mindset. This role is on-site in NYC to support our physical product development.

Qualifications

• Bridge hardware constraints, API limitations, and ML requirements.
• Experience with streaming data, low-latency systems, or sensor processing.
• Coordinate multiple services (transcription, LLM, inference) reliably.

Responsibilities

• Build the backend infrastructure for AI hardware.
• Develop FastAPI services for LLM and device management.
• Collaborate with hardware and data/R&D teams.

Skills

Systems thinking

Real-time data processing

API orchestration

Cross-functional collaboration

Production code delivery

Tools

FastAPI

Supabase

VisPy

Job description

Senior Software Engineer - AI Hardware Backend

Location: NYC - In office.

About Us

Nuha is a research lab based in New York City. We teach companions how to understand, interact with, and respond to humans and the physical world.

Your Role

You'll own the backend infrastructure that powers our AI hardware. This means building the FastAPI services that orchestrate LLMs, audio transcription, sensor data processing, and model inference, connecting our embedded devices to the intelligence that makes them work.

You'll work directly with our hardware team and our data/R&D team to turn sensor streams and voice interactions into intelligent responses.

What You'll Build

• FastAPI services managing WebSocket connections, device sessions, and conversation state
• LLM orchestration (Groq, OpenAI, Modal), conversation memory, prompt engineering, safety guards
• Database architecture (Supabase) - user profiles, interaction history, conversation memories, sensor logs
• Authentication, connection lifecycle, monitoring, and error recovery across production devices
• Process streaming mmWave radar frames and generate tracking visualizations (VisPy)
• Fuse audio, radar, and accelerometer data into coherent context for LLMs
• Build sensor summarization pipelines using VLMs to describe physical space
• Event-driven coordination of audio capture, VAD, TTS playback, and sensor streams

Product Features & Infrastructure

• Ship new interaction features based on customer feedback and product vision
• Build internal tools for debugging device interactions and analyzing sensor data in production
• Design data pipelines for user analytics, interaction logging, and model evaluation
• Performance optimization - reduce latency, improve reliability, handle connection failures gracefully
• Testing infrastructure for hardware-in-the-loop validation and API integration tests
• Iterate on the interaction experience - make conversations feel more natural and responsive

Requirements

• Systems thinking - Bridge hardware constraints, API limitations, and ML requirements
• Real-time data pipelines - Experience with streaming data, low-latency systems, or sensor processing
• API orchestration - Coordinate multiple services (transcription, LLM, inference) reliably
• Cross-functional collaboration - Work with hardware engineers and ML researchers effectively
• Ships production code - Build reliable services that handle real user interactions

Bonus

• Mathematical maturity
• VisPy/graphics experience
• Hardware experience

Team

You'll work alongside engineers doing: RF/mmWave radar design, PCB/enclosure design, embedded client software, ML accelerator R&D, data infrastructure, and manufacturing systems.

Interview Process

• Phone screen
• Technical interview - Standard SWE problem solving and system design
• In-person deep dive - Walk us through a complex technical problem you've solved
• Work trial - Spend time with the team working on a real problem

Location

NYC - In office. We're building a physical product and need to work together in person.

• Any backend/systems projects you've built