Aezion - AI/ML Engineer - Conversational & Generative AI Systems

Description :AI/ML Engineer - Conversational & Generative AI Aezion :Aezion is one of the premiere custom software providers in the United States, living by the adage that our word is our bond.Our Promise is to get it right or make it right.We invest the effort to exceed client expectations across the entire project lifecyclearchitecting, designing, developing, hosting, deploying, maintaining, and supporting intelligent systems.We believe work is ministryan expression of our valuesand we aim to honor commitments through transformative solutions that turn clients into lifelong partners.Role Overview :We are seeking an AI/ML Engineer with expertise in Conversational AI, Generative AI (LLMs), and Speech AI to join our team in building next-generation autonomous AI call systems.This role involves architecting, developing, and deploying real-time AI systems capable of natural human-like conversations, live agent assistance, and end-to-end autonomous call handling at scale.You will be working on technologies spanning speech-to-text (STT), text-to-speech (TTS), LLMs, dialog management, telephony integrations, and real-time AI reasoningdelivering solutions comparable to industry leaders like Cresta, PolyAI, and Google Duplex.Key Responsibilities :Speech AI Development :- Implement low-latency STT (multi-accent, noisy environments) and TTS pipelines using tools like OpenAI Realtime API, Whisper-v3 etc.- Implement voice biometrics for speaker identification and security.Conversational AI & LLM Integration :- Design context-aware dialog management with short-term and long-term memory through autonomous AI Agents.- Fine-tune LLMs (OpenAI GPT-4.5/5, LLaMA, Mistral) for domain-specific reasoning (sales, support, healthcare, etc.- Implement RAG (Retrieval-Augmented Generation) and vector database indexing for real-time contextual responses.Real-Time Call Handling & Telephony Integration :- Integrate with Twilio, Amazon Connect etc. for SIP/VoIP call routing, transcription, and sentiment analysis.- Develop AI-driven call flows with real-time escalation to human agents.- Build "Whisper Mode" for live agent coaching and feedback.AI Feedback & Continuous Learning :- Implement real-time feedback loops for agent assistance, QA scoring, and call analytics.- Use logs for model retraining, context refinement, and error analysis.Security & Compliance :- Ensure PCI-DSS, HIPAA, GDPR/CCPA compliance for all AI-driven interactions.- Implement secure authentication and data governance for AI workflows.Deployment & MLOps :- Develop scalable, containerized AI services using FastAPI, Docker, Kubernetes, and deploy on AWS/GCP/Azure.- Implement MLOps pipelines for continuous integration and delivery of ML models.Key Skills & Experience :- AI/ML Expertise : Strong in LLMs, Agentic AI (CrewAI, LangGraph, MCP etc.), Generative AI (GANs, VAEs), NLP/NLU, speech AI.- Programming : Python (mandatory), FastAPI for microservices, REST APIs.- Frameworks : PyTorch, TensorFlow, Hugging Face Transformers, LangChain, LangGraph.- Vector Databases : Pinecone, PgVector, Qdrant for RAG & semantic search.- Dialog Management : Rasa, Microsoft Orchestration Bot, or custom dialog engines.- Telephony : SIP/VoIP integrations, call routing APIs (Twilio/Amazon Connect).- Streaming AI : Low-latency STT/TTS pipelines with real-time processing.- Cloud & Deployment : Docker, Kubernetes, AWS Lambda, Azure ML, or GCP Vertex AI.- Security : Experience with compliance frameworks for healthcare/finance domains.Nice-to-Have Skills :- LangSmith & LangFuse for LLM observability and debugging.- Knowledge Graphs for enterprise reasoning.- Real-time analytics dashboards using React/Next.js.- Voice Cloning for brand-personalized TTS systems. (ref:hirist.tech)