T
TetraMem - Accelerate The World Tech Stack
Analog computing chips for AI inference at ultra-low power
TetraMem builds analog compute-in-memory hardware for AI workloads, using a deep stack of chip-design tools (SystemVerilog, SPICE, Synopsys, Cadence) and AI frameworks (PyTorch, TensorFlow, JAX, ONNX, TensorRT). The hiring profile is intern-heavy (12 of 17 roles) alongside senior engineers, reflecting a company scaling from R&D into production—projects range from RTL design and chip characterization to neural network deployment and model compression, while pain points center on operationalizing manufacturing and team scaling.
Tech Stack 64 technologies
Core StackPython C++ PyTorch TensorFlow SystemVerilog Verilog SPICE Perl GCC Clang MLIR XLA LLVM FPGA ASIC RTOS Zephyr FreeRTOS Synopsys Cadence Mentor Graphics Cadence Virtuoso Google Identity JAX C/C++ ONNX TensorRT TensorFlow Lite ncnn Core ML+32 more
AdoptingDjango
What TetraMem - Accelerate The World Is Building
◆Challenges
- Manual inventory tracking
- Asset visibility
- System implementation
- Neural network model efficiency
- Moving from r&d into commercialization
- Building and scaling high-performing technical teams
▲Active Projects
- Lab asset tracking tool
- Erp implementation
- Product demo setup
- Benchmark/showroom projects
- Reram characterization
- Reliability testing
- Device performance improvement
- Ai accelerator rtl design and verification
- Neural network deployment on compute-in-memory chips
- Model compression and conversion for analog compute-in-memory hardware
Hiring Activity
15 roles · 15 in 30d
Department
Seniority
Company intelligence
