Technical Architect

Responsibilities

• Core AI/ML Fundamentals
• Solid understanding of AI/ML concepts including:
• Classification, regression, neural networks
• OCR and transcription systems
• Audio/Video processing and multimodal learning
• OCR, Transcription & Audio/Video Intelligence
• Implement specialized models for:
• Highaccuracy document OCR
• Realtime audio transcription
• Architect deep learning pipelines for audio/video analysis and generation.
• Integrate multimodal models (e.g., LLaVA, Whisper) into broader GenAI systems.
• Generative AI & LLM Expertise
• Strong understanding of:
• Generative AI techniques
• Transformer architectures
• RAG (Retrieval-Augmented Generation) pipelines
• Modern LLM ecosystems
• Hands-on experience with:
• LLM parameter handling and model selection
• Scaling strategies and performance optimization
• Expertise in:
• Prompt engineering and instruction tuning
• Prompt tuning and optimization for high-quality outputs
• Familiarity with evaluation frameworks covering:
• Quality, grounding, accuracy, safety
• Latency and cost analysis
• Governance and compliance requirements
• Agentic AI Systems
• Experience designing and building agentic AI systems including:
• Multiagent orchestration
• Tooluse workflows
• Autonomous task execution
• Design longterm memory architectures:
• Vector-based memory systems
• Graph-based memory for complex, persistent context
• Model Training, FineTuning & Optimization
• Knowledge of finetuning approaches:
• LoRA, QLoRA, supervised fine-tuning
• Experience with model compression techniques:
• Quantization, distillation
• Familiarity with performance-level tooling:
• CUDA, Triton, or specialized custom kernels
• Design AI systems capable of efficiently handling and routing multiple user requests simultaneously, including:
• Scalable request handling
• Loadbalanced inference
• Multitenant model utilization
• Caching and prioritization strategies
• Infrastructure, Pipelines & Deployment
• Oversee:
• Data pipeline integration
• Training workflows
• ML CI/CD processes
• Strong understanding of:
• GPU/compute requirements
• Costefficient deployment strategies
• Experience designing and managing production-grade inference servers using:
• vLLM
• Text Generation Inference (TGI)
• SGLang
• Ability to collaborate with engineering teams to integrate LLMs into production systems:
• APIs, microservices, cloud architectures
• Research, Evaluation & Continuous Innovation
• Stay current with advancements in AI, ML, and LLM ecosystems.
• Evaluate new tools, frameworks, and platform technologies to continuously enhance system architecture.