Block or report aidyai
Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.
You must be logged in to block users.
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abuseIDARA SAMUEL OSU — Machine Learning Engineer
Speech AI • Low-Latency Inference • Agentic Systems • Production MLOps
CORE STACK & TOOLING
Inference Engineering:
TensorRT-LLM vLLM SGLang CTranslate2 KV Caching Quantization In-Flight Batching
ML, NLP & Speech:
PyTorch Hugging Face LLMs Instruction Tuning NMT (Meta NLLB) TTS / STT Whisper Pipecat WebRTC / VAD
Agentic Frameworks:
LangGraph LangChain Multi-Agent Orchestration Tool / Function Calling
Cloud & MLOps Infrastructure:
AWS SageMaker Model Registry Multi-Model Endpoints Model Monitor Modal Docker CI/CD Supabase Python
TECHNICAL FOCUS & CAPABILITIES
|
Low-Latency Serving Stack Compiling open-weight LLMs, Whisper, NMT, and TTS/STT models via TensorRT-LLM and CTranslate2; implementing KV caching, quantization, and in-flight batching for sub-second execution. |
|
Full-Duplex Interactive Voice Architecting real-time speech-to-speech pipelines using Pipecat with WebRTC, VAD-driven turn-taking, and barge-in support for conversational voice agents. |
|
Agentic Orchestration Designing stateful multi-tool routing environments and multi-agent coordination architectures using LangChain and LangGraph with function calling. |
|
Cloud-Native MLOps & Production Pipelines Building automated SageMaker Pipelines (preprocess, train, evaluate, register), multi-model endpoints, serverless inference layers, and automated CI/CD deployments with Docker. |
|
Data Curation & Multilingual AI Architecting raw dataset collection pipelines across multi-thousand-hour speech and text corpora for low-resource and zero-digital-footprint languages. |
📬 Contact: idaraosusamuel@gmail.com | LinkedIn
Pinned Loading
If the problem persists, check the GitHub status page or contact support.