H2B Group

Technical Lead — Real-Time Voice Translation Platform

Brak informacji o wynagrodzeniu

C-Level / ManagerZdalnie·GdańskFull-time·B2B

#455220·Dodano 26 dni temu·1

Źródło: justjoin.it

Aplikuj teraz

Tech Stack / Keywords

PythonWebRTCReact Native

Podsumowanie oferty

The role is for a Technical Lead responsible for designing and developing real-time audio processing systems. The project is an innovative mobile application for live voice translation enabling natural conversation between users speaking different languages. The Technical Lead will oversee the initial discovery sprint, define architecture, coordinate with clients, and plan subsequent development phases.

Wymagania

Proven production experience in real-time audio processing systems, particularly streaming voice pipelines; batch transcription or request/response AI services experience insufficient.
Practical experience in at least two of the three areas: Streaming Speech-to-Text (STT), streaming or incremental Machine Translation (MT), streaming Text-to-Speech (TTS).
Production experience with WebRTC, including audio transmission behavior under unstable cellular network conditions.
Experience in instrumentation and latency measurement: endpoint measurements, median and p95 analysis, latency distribution interpretation, outlier analysis.
Ability to evaluate technology providers for compliance, especially EU data hosting and zero data retention guarantees.
Very good knowledge of Python for backend development.
Familiarity with frameworks for orchestrating streaming audio pipelines, e.g., Pipecat or equivalent solutions.
Experience at Technical Lead or Architect level, independently making, documenting, and defending technical decisions.
Ability to analyze trade-offs between latency, quality, costs, scalability, and data privacy.
Very good English proficiency in speaking and writing to conduct client meetings and prepare precise technical documentation.
Availability to start soon and full commitment for the first two project weeks.

Obowiązki

Designing solution architecture: analyze and choose optimal approaches for real-time speech translation including STT → MT → TTS versus end-to-end speech-to-speech models.
Designing streaming audio processing: develop architecture enabling playback of translated speech before the original utterance ends.
Analyzing and selecting AI providers: evaluate Speech-to-Text, Machine Translation, Text-to-Speech, and WebRTC solutions focusing on streaming and compliance.
Defining latency budgets: identify and measure delays at each pipeline stage, determine expected values and variability, and identify main delay sources.
Evaluating speech synthesis quality: compare TTS variants by voice naturalness, speaker similarity, streaming efficiency, multilingual support, and cost.
Designing testing methodology: define criteria for transcription and translation quality, measure time-to-first-audio including percentile and outlier analyses.
Setting Alpha stage criteria: prepare measurable acceptance criteria and pass/fail thresholds agreed with the client before implementation.
Analyzing solution costs: create cost models for processing audio minutes under different usage scenarios.
Assessing on-device processing feasibility: analyze potential to migrate pipeline components to mobile devices.
Preparing technical foundations: configure repository, CI/CD, app skeleton, and basic authentication mechanisms.
Preparing architectural documentation: produce recommendations, technical justifications, comparative analyses, and architectural decision records in English.
Collaborating with client: lead technical discussions, present recommendations, and justify decisions based on data and technical analysis.
Planning next phase: propose scope for Alpha version, estimations, team structure, and preliminary budget.

Benefity

Immediate start possibility; client can wait up to 2 weeks.
Stable B2B contract for minimum 3 months with potential long-term extension.
100% remote work (no office or client site visits required).
Fast recruitment process: single interview with developers and HR, feedback within about a week.

Inne informacje

The data controller is H2B Group Sp. z o.o., located at Ul. Abrahama 1A, 80-337 Gdańsk. Applicants have rights regarding personal data access, correction, deletion, processing restrictions, objections, data portability, and complaints to supervisory authorities. Personal data is processed for recruitment purposes as mandated by law or by consent. Refusal to provide mandatory data may prevent recruitment. Data will be retained until the end of recruitment and statute of limitations periods. Consent for future recruitment participation can be withdrawn at any time.

H2B Group

53 aktywne oferty

Zobacz wszystkie oferty

Aplikuj teraz