The main goal of the Unbreakable Ukraine Foundation — is to provide Ukrainian refugee children with uninterrupted access to the formal and non-formal education According to UNICEF, more than 800,000 children came to Poland, and only 190,000 of them go to Polish schools. The rest study at home in their Ukrainian schools online. We offer everything the same as the Ukrainian school in Ukraine, but offline and in Poland. "The First Ukrainian School in Poland" was established by the Foundation in March 2022 by Viktoriia Gnap, an ideologist and president of the Foundation. On August 24 "The First Ukrainian School in Poland" received a Ukrainian state license #008137/22 to conduct educational activities at the level of preschool education, complete general secondary education, in the field of culture, and was registered as "Kyiv Private Lyceum "First Ukrainian School" from the Foundation "Unbreakable Ukraine". The school now continues its work in three cities of Poland in the 2022-2023 academic year: Warsaw, Wroclaw, and Krakow. Education is carried out in accordance with the recommendations of the Ukrainian Ministry of Education and Science, by Ukrainian teachers. We aim to provide students with free education and everything they need to study: books, stationery, backpacks, and even sports equipment. We provide children with free lunches daily. The project is implemented with the financial support of the United Nations Children’s Fund (UNICEF) and International Organization Save the Children. Everyone can help Ukrainian children by donating to the Foundation "Unbreakable Ukraine" or by financing our projects. Our most urgent need is the funds for teachers' salaries, rental of premises, educational materials and books, and nutrition of children. Please call us, write to us on social networks or at the contacts listed on the site, and we will contact you and tell you what opportunities there are now.
-
· 16 views · 2 applications · 27d
Senior Streaming ASR/TTS Engineer
Part-time · Full Remote · Worldwide · Product · 1 year of experience · English - B2Project Overview: We are building a real-time voice agent (voice→text→voice) integrated with our proprietary software stack. We need a senior part-time engineer for a few months to select, deploy, optimize, and fine-tune open-source ASR and TTS models for...Project Overview:
We are building a real-time voice agent (voice→text→voice) integrated with our proprietary software stack. We need a senior part-time engineer for a few months to select, deploy, optimize, and fine-tune open-source ASR and TTS models for ultra-low-latency streaming and production-grade stability.Primary languages: English, Ukrainian, Polish, Russian, German, Spanish, French, Portuguese (+ others as needed). A critical requirement is high-quality Slavic stress/pronunciation (especially UA/PL/RU) and robust handling of mixed-language inputs (code-switching).
Must-have:
- Proven production experience with streaming ASR and/or streaming TTS (real demos/links required).
- Strong real-time audio knowledge: VAD, endpointing, chunking, partial/final hypotheses, buffering/jitter, barge-in (interrupt handling).
- Practical approach to text normalization (numbers/dates/currency/abbreviations) and Slavic stress/pronunciation (dictionaries/rules/G2P/SSML).
- Linux + Docker, reproducible deployments, clean documentation.
Responsibilities:
- Select and justify an open-source ASR/TTS stack for streaming and multilingual performance.
- Implement streaming ASR: partial/final, VAD + endpointing, stability, metrics.
- Implement streaming TTS: low time-to-first-audio, stable chunked audio streaming (no dropouts), prosody control.
- Build a text normalization + pronunciation/stress layer with tests (focus on UA/PL/RU).
- Create training/fine-tuning pipelines on our private data + evaluation/regression.
- Integrate into our agent pipeline (voice→text→LLM→text→voice), including barge-in and turn-taking.
- Deliver Dockerized services + runbook (deploy/train/evaluate/debug).
Deliverables:
- Dockerized Streaming ASR service + documented API.
- Dockerized Streaming TTS service + chunked audio streaming.
- Text normalization + pronunciation/stress module (configurable dictionaries/rules, tests).
- Reproducible ASR/TTS training/fine-tuning pipelines.
- Evaluation + regression suite (latency + WER/CER + Slavic stress checklist).
- End-to-end integration demo with our agent + instructions.
- Documentation (architecture notes + runbook).
Acceptance Criteria (measured):
- Latency: ASR time-to-first-partial; TTS time-to-first-audio; end-to-end “user stops → agent starts speaking”.
- ASR quality: WER/CER on multilingual test sets (EN/UA/PL/RU/DE/ES/FR/PT).
- Pronunciation/stress: pass rate on a curated Slavic stress/pronunciation checklist (UA/PL/RU).
- Stability: reliable barge-in and turn-taking in real-time streaming.
- Reproducibility: Docker deployment + repeatable training runs.