Sr Architect - Agentic AI

· Bachelor’s or Master’s degree in Computer Science, AI/ML, Data Science, or related field

· Advanced coursework or certifications in machine learning, deep learning, or NLP

· Strong mathematical and statistical foundation

Experience

· 6-10 years of experience in AI/ML/Deep Learning model development

· 2-3 years Hands-on experience with LLMs, NLP, or speech/voice AI systems

· 3-5 years of experience deploying AI solutions in production environments

Primary (Must have skills)

· 5+ years of experience in Python, PyTorch, TensorFlow, or similar frameworks

· 3-5 years of experience designing, training, and fine-tuning large language models or AI models, speech/voice AI systems, Realtime voice pipeline

· 5 years of experience in cloud AI platforms (AWS Sagemaker, Azure ML, Google Cloud Platform AI)

· 2+ years of experience in Agentic AI frameworks such as LangChain, LangGraph, A2A, MCP, and Multi agent orchestration

· 2-3 years of experience Familiarity with model evaluation metrics, bias detection, and optimization

· 5 years of experience in integrating AI models into applications via APIs or pipelines

· 2-3 years’ experience in Azure services - Azure AI Speech and Translator, Azure OpenAI, AI Search, plus Container Apps/AKS, API Management, Event Hubs, Key Vault, and Application Insights

Key technical skills:

· Design, develop, and deploy advanced AI models to meet business requirements

· Collaborate with data engineers, LLM Ops, and software teams for end-to-end solutions

· Evaluate model performance, fine-tune, and ensure scalability and reliability

· Research emerging AI/ML technologies and propose innovative solutions

· Mentor junior AI engineers and review code/models for best practices

Secondary Skills:

· Advanced LLM techniques, prompt engineering, and fine-tuning strategies

· AI ethics, fairness, and responsible AI deployment

· Continuous learning on emerging AI frameworks and architectures

· LLMOps - layered evals (WER, entity F1, COMET), CI regression gates, tracing, cost-per-minute telemetry, model routing, and drift detection.