Hiring Intern/Junior AI Engineer (Speech AI)

8 September, 2026

Careers

Trivita AI is hiring an Intern / Junior AI Engineer (Speech AI) to research and develop ASR, TTS, speech enhancement, and voice technologies.

Job description

Research and develop Speech AI solutions related to:
- Speech-to-Text (ASR)
- Text-to-Speech (TTS)
- Speech Enhancement / Denoising
- Voice Activity Detection (VAD)
- Speaker Diarization / Speaker Recognition
Build pipelines for audio data processing and evaluation.
Preprocess, label, validate, and analyze speech data.
Experiment with open-source or commercial models such as Whisper, wav2vec 2.0, MMS, NVIDIA NeMo, CosyVoice, and similar models.
Fine-tune and optimize models for Vietnamese and real-world use cases.
Evaluate model quality using metrics such as WER, CER, MOS, latency, and throughput.
Collaborate with Backend and MLOps teams to deploy models as APIs or production services.
Read technical documentation and research papers, and stay up to date with emerging Speech AI technologies.

Third-year or final-year students, or recent graduates in Computer Science, Artificial Intelligence, Data Science, Electrical/Computer Engineering, or related fields.
Basic knowledge of Python, Machine Learning, and Deep Learning.
Knowledge of signal processing or audio processing is a plus.
Experience using frameworks such as PyTorch or TensorFlow.
Strong self-learning mindset with the initiative to experiment and read technical documentation.

6 months to 2 years of experience in AI/ML or Speech AI.
Proficiency in Python and at least one Deep Learning framework, preferably PyTorch.
Understanding of one or more areas such as ASR/TTS, Audio Classification, Speech Enhancement, Voice Conversion, or Speaker Diarization.
Experience working with audio datasets, feature extraction, and model evaluation.
Familiarity with Git, Docker, and Linux.
Experience deploying models as APIs is a plus.

Personal projects or research experience related to Speech AI.
Experience with Whisper, wav2vec 2.0, NVIDIA NeMo, Hugging Face, or ONNX.
Understanding of Mel-spectrograms, MFCC, FFT, sampling rates, and noise reduction.
Experience with inference optimization techniques such as quantization, batching, streaming, or GPU serving.
Published articles, research papers, GitHub projects, or contributions to open-source projects.

Work on real-world Speech AI products serving actual users.
Receive guidance from engineers experienced in AI and production systems.
Work with Vietnamese speech data and large-scale Speech AI problems.
Gain hands-on experience in model training, evaluation, deployment, and performance optimization.
Work in an environment that encourages research, experimentation, and new ideas.
Opportunity to become a full-time employee after the internship.