S
Sesame is building custom embedded operating systems and firmware to run large language models on consumer wearables in real time. The stack—Linux, RTOS (Zephyr, FreeRTOS, QNX), C/C++, PyTorch, vLLM, SGLang—reflects a hardware-software integration challenge: they're not just training models but optimizing them for power, latency, and memory on constrained devices. Engineering dominates hiring (40 of 71 roles), with 50 senior-level positions, signaling deep technical depth required to bridge firmware, AI inference, and manufacturing at scale.
Tech Stack 55 technologies
Core StackLinux Python PyTorch AWS Kubernetes RTOS C/C++ AOSP Zephyr FreeRTOS DSP Bluetooth WiFi Ethernet USB Qualcomm QNX Assembly JTAG GDB vLLM SGLang GCP Azure Ray iOS iOS SDK Apple WebRTC WebSockets+25 more
What Sesame Is Building
◆Challenges
- Scaling training and inference for llm workloads
- Scalability challenges across domains
- Improving application reliability
- Improving developer productivity
- Tight timelines
- Improving manufacturing efficiency
- Reducing model initialization times
- Optimizing algorithms for power, latency, and memory footprint
- Improving manufacturing test coverage
- Gap between research and production
▲Active Projects
- Custom embedded operating system for ai-driven real-time consumer applications
- React-based web tools for data manipulation
- Turbocharge serving layer
- Wearable consumer product development
- Manufacturing test strategy for new product introductions
- Test fixture and equipment design
- Real-time audio capture and render systems
- Firmware development workflow improvement
- Scaling llm training and inference
- End-to-end hardware development process
Hiring Activity
70 roles · 30 in 30d
Department
Seniority
Company intelligence
