Machine Learning Engineer, Automatic Speech Recognition
Machine learning engineer owning the ASR improvement loop for Kiefer Tech's Sophea Meet meeting-transcription product: training/fine-tuning speech recognition models on real meeting audio, building benchmarks and error analyses, and improving production inference. Python with PyTorch/JAX/TensorFlow; English focus first, then Greek.
- Compensation
- Not disclosed
- Location
- Athens, GR
- Work setup
- Not specified
- Seniority
- Not specified
- Employment
- Full time
- Experience
- 3+ years
- Category
- Ml / AI
- Posted
Skills
Must have
- 3+ years of relevant ML or research engineering experience with technical ownership and measurable results
- Strong hands-on ASR or speech ML experience, including a shipped or materially improved speech recognition system
- Experience training, fine-tuning, or adapting deep learning models on real-world audio data
- Strong evaluation skills: WER/CER, benchmark design, error analysis, regression testing
- Strong Python and hands-on experience with PyTorch, JAX, TensorFlow, or similar; solid software engineering skills
- Experience deploying, monitoring, and maintaining production ML systems
Nice to have
- Multilingual ASR, accent robustness, domain adaptation, or diarization
- Streaming speech recognition and low-latency inference
- MLOps, evaluation, or data pipelines for speech systems
- Experience with meeting transcription, voice agents, or enterprise speech products
Full description
ML Engineer — ASR
Department: Technology
Location: Athens
Employment Type: FullTime
About the company:
KIEFER is building Greece's integrated AI ecosystem. From renewable energy infrastructure and AI systems to robotics and enterprise applications, we connect the technologies that power Europe's intelligent future. Founded in 2014, KIEFER has delivered 600MW+ of energy projects and is now developing sovereign AI infrastructure, enterprise AI products and physical AI systems for Greece and Southeast Europe.
About the role
We are looking for an ML Engineer to improve automatic speech recognition (ASR) for Sophea Meet. You will work with real meeting audio and own the improvement loop from data and model experiments through evaluation and production performance.
The initial language focus is English. Additional languages, starting with Greek, will be prioritized with Product based on usage and product needs.
What you will do
- Train, fine-tune, adapt, and evaluate speech recognition models based on measured product needs.
- Build representative benchmarks and regression tests for real meeting audio.
- Prepare speech datasets and investigate difficult cases, including noise, overlapping speakers, long sessions, and specialized terminology.
- Use error analysis to decide which model, data, or pipeline changes will improve transcription quality.
- Improve inference across accuracy, latency, throughput, reliability, and cost.
- Deploy and monitor model versions with Engineering, while working with Product to set quality goals and prioritize improvements.
What you will need
- 3+ years of relevant ML or research engineering experience, with clear technical ownership and measurable results.
- Strong hands-on ASR or speech ML experience, including a speech recognition system you shipped or materially improved.
- Experience training, fine-tuning, or adapting deep learning models using real-world audio data.
- Strong evaluation skills, including WER/CER, benchmark design, error analysis, and regression testing.
- Strong Python, hands-on experience with PyTorch, JAX, TensorFlow, or a similar framework, and solid software engineering skills.
- Experience deploying, monitoring, and maintaining production ML systems.
Nice to have
- Multilingual ASR, accent robustness, domain adaptation, or diarization.