Ustym K.

AI Audio Engineer specializing in ASR, TTS, voice cloning, and generative music systems. I build and deploy production-ready audio AI pipelines — from speech recognition and synthesis to real-time music generation and audio intelligence. Every system is designed for low latency, edge deployment, and real-world scalability. I’m Ustym, an AI & Machine Learning expert and tech lead specializing in real-time speech, music, and audio processing. I help startups and R&D teams design, prototype, and launch AI-powered audio solutions that are robust, scalable, and ready for production. Whether you’re building a TTS voice assistant, enhancing voice quality, or generating music with style conditioning — I’ve got you covered. 🧰 WHAT I OFFER 🔹 Design and train ASR, TTS, voice cloning, and audio generation models 🔹 Build production-ready pipelines for speech, music, and sound AI 🔹 Optimize models for edge devices (ONNX, TensorRT, TFLite) 🔹 Deliver full-cycle ML development: from research to deployment 🔨 WHAT I BUILD I work across the full pipeline: research, data processing, model training, evaluation, and deployment. My solutions are designed for real-time performance, low latency, and production scalability. 🗣️ AI Speech Systems - ASR (speech-to-text), TTS (text-to-speech), voice cloning, speech enhancement - Speaker identification, phoneme recognition, prosody analysis - Robust to accents, emotions, background noise, and multilingual input 🎵 AI Music & Audio Generation - Generative music with mood/style/genre control, BPM/key detection - Instrument transcription, music captioning, stem separation, audio-to-MIDI - Trusted by music tech platforms and sound design teams 🔉 Audio Intelligence & Sound Analysis - Audio event detection, acoustic scene classification, similarity search - Text-to-audio generation, real-time audio understanding - Applied in smart environments, accessibility, health tech 🛠️ TECH STACK 🧠 AI Frameworks: PyTorch, TensorFlow, Keras, JAX, Hugging Face Transformers, Diffusers 🎙️ Speech & Audio Processing: Whisper, NVIDIA NeMo, SpeechBrain, Kaldi, ESPnet, Coqui TTS Librosa, Torchaudio, FFmpeg, SoX, PyDub, PyAudio, PortAudio 🎧 Generative Audio Models: StyleTTS, Bark, ElevenLabs, Descript Overdub, Magenta, MusicLM, AudioLM, Riffusion, DDSP ⚙️ Model Deployment: ONNX, TensorRT, TorchScript, TFLite, Docker, FastAPI ☁️ Cloud & MLOps: AWS (SageMaker, Lambda, EC2), GCP, Azure, MLflow, DVC, GitHub Actions 👨‍💻 Programming & Data Tools: Python, C++, JavaScript, NumPy, Pandas, SciPy, SQL, Git 🎯 INDUSTRIES I SUPPORT ✔️ Media & Entertainment – multilingual dubbing, personalized streaming, voice synthesis ✔️ Music Tech – AI tools for composition, tagging, sound design ✔️ Gaming & XR – dynamic sound, NPC voices, real-time audio events ✔️ Healthcare – assistive tools, voice diagnostics, accessibility ✔️ EdTech – pronunciation feedback, speech tutoring, language training ✔️ Smart Devices & Security – voice authentication, sound-based alerts ✔️ Customer Experience – transcription, voice agents, sentiment detection ✔️ Video & Film – dubbing, adaptive soundtracks, intelligent editing 💡 WHY ME 🔹 Deep Audio AI Focus – I specialize 100% in speech, music, and sound AI 🔹 Research-Driven Development – I follow the latest ML innovations and apply them fast 🔹 Production-Oriented Mindset – Models are built for deployment, not just demos 🔹 Business Alignment – My work supports real product outcomes and ROI I lead a team of AI/ML audio engineers at It-Jim, where we’ve built systems for voice tech startups, creative tools, medtech products, and smart environments. We bring in strong expertise in ASR, TTS, audio DSP, and full-stack ML delivery. Let’s talk if you need a hands-on expert to build or improve your Audio AI system. Whether it’s voice, music, or sound recognition — I’ll help you go from idea to real-world solution. Let’s build something amazing with sound!