Skip to content
VoxNexus
HOME / ABOUT
ABOUT VOXNEXUS

Leading voice AI, on-device.
For the hardest languages.

We are a founding team in speech algorithms and engineering, committed to bringing state-of-the-art speech and translation capabilities truly on-device for emerging markets — and to connecting every corner of the world through voice.

Open positions ↓ Contact us ↗
01 / JOIN US

Become a pioneer
of on-device AI.

We are building a founding core team in speech algorithms and engineering, focused on tackling the world's “hardest” languages — committed to bringing leading speech and translation capabilities truly on-device for emerging markets. Join us as a fellow traveler, and become a pioneer of on-device AI.

01 / STRUCTURE

Efficient organizational structure

We maintain a flat organizational design, committed to building a focused, undisturbed R&D environment. Problem-oriented, we emphasize rapid response and efficient decision-making.

02 / CULTURE

Autonomy and innovation culture

We champion a highly inclusive culture of innovation that encourages employees to think independently, break from convention, and push the boundaries of technology. There is no ceiling on business growth or personal development — everyone is given ample autonomy to unlock their full potential.

03 / GLOBAL TEAM

A global team with a global outlook

We have established legal entities in both Singapore and the United States, with our business deeply serving clients worldwide. This is an elite team with an international perspective and a fusion of diverse cultures.

02 / OPEN POSITIONS

Two founding roles are open today.

Select a role to see the full description.

Responsibilities

  1. 01 Responsible for model training and inference acceleration optimization, continuously improving training efficiency, inference performance, resource utilization, and system stability.
  2. 02 Responsible for the engineering implementation, system integration, and performance optimization of algorithms related to speech signal processing, automatic speech recognition (ASR), and text-to-speech (TTS).
  3. 03 Responsible for the full-pipeline engineering delivery of speech algorithms from training to deployment, including model conversion, graph optimization, operator adaptation, runtime integration, inference pipeline integration, and performance validation.
  4. 04 Lead model acceleration and optimization work, including quantization, pruning, distillation, operator fusion, parallel acceleration, and mixed-precision optimization, balancing accuracy, latency, throughput, memory, and power consumption.
  5. 05 Carry out model deployment and performance tuning for on-device, embedded, or heterogeneous hardware platforms, improving the usability and delivery efficiency of algorithms in real business scenarios.
  6. 06 Responsible for diagnosing and resolving complex issues, including training efficiency bottlenecks, abnormal inference latency, accuracy degradation, operator compatibility, excessive memory usage, and insufficient real-time performance.
  7. 07 Collaborate with algorithm, platform, embedded, software, and product teams to drive the joint debugging, deployment, and mass-production delivery of speech capabilities in actual products.
  8. 08 Participate in inference engine architecture design, core module development, and technical proposal reviews, and help build out platform capabilities and consolidate engineering standards.
  9. 09 Mentor junior and mid-level engineers in model deployment, performance tuning, and engineering implementation, improving the team's overall engineering capability.

Qualifications

  1. 01 Bachelor's degree or above in Computer Science, Electronics, Communications, Automation, Artificial Intelligence, Signal Processing, or a related field.
  2. 02 5+ years of experience in algorithm engineering, inference engines, or model deployment optimization, with a track record of complete project delivery and mass production.
  3. 03 Experience in both model training acceleration optimization and model deployment/inference acceleration optimization, with the ability to independently complete performance analysis, bottleneck identification, and optimization loops.
  4. 04 Experience in the engineering implementation and optimization of speech signal processing, ASR, and TTS, with the ability to independently drive algorithms from proposal to product deployment.
  5. 05 Familiar with deep learning model training, deployment, and inference workflows, with experience in model compression, quantization, acceleration, and on-device deployment.
  6. 06 Familiar with one or more mainstream inference frameworks or deployment tools, such as ONNX Runtime, TensorRT, TFLite, NCNN, MNN, SNPE, etc.
  7. 07 Proficient in C/C++ and Python development, with solid engineering implementation skills, systems-level problem analysis skills, and performance optimization skills.
  8. 08 Strong cross-team communication and project-driving abilities, able to serve as the owner or technical lead of core modules.

Preferred Qualifications

  1. 01 Experience delivering on-device AI, embedded AI, mobile AI, or real-time voice interaction products.
  2. 02 Experience with model deployment and performance optimization on DSP, NPU, GPU, or other heterogeneous computing platforms.
  3. 03 Experience with AI inference deployment and acceleration optimization on Qualcomm, MTK, or other SoC platforms.
  4. 04 Experience developing proprietary inference engines/runtimes, graph execution optimization, operator development, or compiler optimization.
  5. 05 Experience with Android HAL, Framework, or application-layer development; candidates able to coordinate delivery across algorithms, underlying systems, and upper-layer business logic are preferred.
  6. 06 Experience deploying large-model speech capabilities, streaming inference, low-power optimization, or building mass-production stability.
JOIN VOXNEXUS

Don't see your role?
[ say hello ]

We'd still love to hear from you. Reach out and tell us how you'd help bring voice AI on-device.

HR@VOXNEXUS.AI ↗