AI Architect Full Time

Join Delphi – Where Innovation meets transformation

At Delphi, we believe in creating an environment where our people thrive. Our hybrid work model empowers you to choose where you work—whether it’s from the office, your home, or a mix of both— so you can prioritize what matters most. We are committed to supporting your personal goals, family, and overall well- being while driving transformative results for our clients.

We welcome exceptional talent from anywhere across the globe. Interviews and onboarding are conducted virtually, reflecting our digital-first mindset.

Rooted in the region, we specialize in delivering tailored, impactful solutions in Data, Advanced Analytics and AI, Infrastructure, Cloud Security, and Application Modernization. Whether it’s enabling predictive analytics , transforming operations with automation, or driving customer engagement with intelligent platforms, we are the trusted partner for organizations ready to embrace a smarter, more efficient future.

About the Role

Senior AI/ML Technical Architect – Generative AI & Agentic Systems

Location: India (Remote)

We are seeking a Senior AI/ML Solution Architect with deep expertise in Generative AI and agentic systems to lead the architecture and delivery of enterprise-scale AI solutions. This role requires a strong combination of hands-on technical capability across Large Language Models (LLMs) and Small Language Models (SLMs), along with the architectural leadership to design, integrate, and deploy AI systems across cloud, edge, and on-premises environments.

The successful candidate will architect scalable agentic platforms, build advanced RAG and fine-tuning pipelines, and design robust integration frameworks connecting AI services with enterprise applications. This role sits at the core of AI transformation initiatives, balancing cutting-edge innovation with production-grade engineering, performance optimization, and security-first design.

Experience Requirements

• 8+ years of experience in software engineering, data, or AI systems
• 2+ years of hands-on experience with Generative AI and LLM-based solutions
• 4+ years of experience designing and architecting enterprise-scale platforms and distributed systems

Key Responsibilities

Architecture & System Design

• Architect scalable agentic systems using advanced LLM and SLM capabilities
• Design multi-agent orchestration frameworks for complex, automated workflows
• Define context, memory, and state-management architectures for persistent agent interactions
• Implement Model Context Protocol (MCP)–based integrations across enterprise services

AI & Platform Implementation

• Design and optimize Retrieval-Augmented Generation (RAG) architectures
• Build agent solutions using LangChain, LangGraph, Semantic Kernel, Agno, and custom frameworks
• Architect and deploy model inference pipelines across cloud, edge, and on-prem environments
• Develop fine-tuning strategies for LLMs and SLMs, including domain and task specialization
• Lead model compression, quantization, and performance-optimization initiatives

Integration & Enterprise Connectivity

• Architect secure REST, gRPC, and GraphQL APIs for AI platform services
• Design event-driven architectures using message buses and webhooks
• Implement authentication and authorization systems (SSO, OIDC, token-based security)
• Build and govern enterprise connectors (Slack, Jira, Salesforce, ERP/CRM platforms)

Data, Models & Evaluation

• Design data preprocessing and governance pipelines (cleaning, deduplication, PII handling)
• Architect embedding generation, re-indexing, and retrieval optimization workflows
• Define chunking, windowing, and content processing strategies
• Establish model evaluation, benchmarking, and selection frameworks

Required Technical Skills

Core AI & GenAI

• Deep experience with GPT-4, Claude, LLaMA, and enterprise/open-source LLM ecosystems
• Expertise deploying and optimizing SLMs (Phi-3, Gemma, TinyLlama)
• Advanced agent frameworks: LangChain, LangGraph, Semantic Kernel, Agno
• RAG systems, vector databases, semantic and hybrid retrieval

Fine-Tuning & Model Optimization

• Parameter-efficient tuning: LoRA, QLoRA, DoRA, AdaLoRA
• Prompt-tuning, adapters, prefix tuning, P-tuning v2
• RLHF / RLAIF pipelines
• Compression techniques: quantization (INT8/INT4, GPTQ, AWQ, GGML), pruning, distillation

Deployment & Performance

• Multi-environment deployment (cloud, edge, on-prem)
• Autoscaling, rate