True Tech Professionals
Website:
truetechpro.io
Job details:
Job Title: Principal AI/ML Engineer – Multi-Agent AI Systems
Location: Gurugram, India (On-site)
Job Type: Full-time
Reports To: AI Solutions Architect / Engineering Leadership
Team: Lead AI/ML Engineering Team (2–4 Engineers)
About the Role
We are looking for a highly experienced Principal AI/ML Engineer to lead the design and development of enterprise-grade, production-ready multi-agent AI systems powered by Large Language Models (LLMs). This is a hands-on technical leadership role focused on building scalable agentic architectures, intelligent orchestration frameworks, retrieval-augmented generation (RAG) systems, evaluation pipelines, and AI governance capabilities.
You will define the architecture for AI agents, establish engineering best practices, mentor a small team of AI engineers, and drive the technical vision for next-generation AI applications that require reliability, explainability, security, and enterprise-scale performance.
Key Responsibilities
- Design and implement scalable multi-agent AI architectures for complex enterprise workflows.
- Build orchestration layers that coordinate multiple AI agents, tools, and services efficiently.
- Design intelligent reasoning agents for research, retrieval, planning, content generation, and structured decision-making.
- Develop Retrieval-Augmented Generation (RAG) pipelines using vector databases, hybrid search, embeddings, and reranking techniques.
- Architect AI governance frameworks including guardrails, auditability, provenance, authorization, and secure data isolation.
- Define agent communication protocols, context management, retry strategies, and failure handling mechanisms.
- Evaluate and optimize LLM performance, latency, token usage, and operational cost.
- Define model selection strategies, balancing Small Language Models (SLMs) and Large Language Models (LLMs) based on workload requirements.
- Design model deployment, serving, versioning, and CI/CD pipelines for production AI systems.
- Implement evaluation frameworks to measure hallucination, retrieval quality, factual accuracy, and structured output reliability.
- Build personalization and adaptation strategies using prompt engineering, parameter-efficient fine-tuning (LoRA/PEFT), and lightweight model customization where appropriate.
- Collaborate with product, engineering, and QA teams to establish quality standards and production readiness.
- Mentor AI engineers, conduct architecture reviews, and provide technical leadership across projects.
Required Qualifications
- 10+ years of experience in software engineering, machine learning, or AI systems.
- At least 4 years of hands-on experience building production-grade LLM or agentic AI applications.
- Strong expertise in Python and asynchronous programming.
- Deep experience with multi-agent orchestration frameworks such as LangGraph, LangChain, or similar platforms.
- Extensive knowledge of Retrieval-Augmented Generation (RAG), embeddings, vector databases, hybrid retrieval, reranking, and retrieval evaluation.
- Experience designing enterprise AI systems with governance, security, auditability, and data isolation.
- Strong understanding of LLM evaluation techniques, hallucination mitigation, structured output validation, and guardrails.
- Experience deploying, serving, optimizing, and monitoring LLMs in production environments.
- Knowledge of model adaptation techniques including LoRA, PEFT, prompt tuning, or supervised fine-tuning.
- Experience with model lifecycle management, CI/CD pipelines, versioning, and MLOps best practices.
- Practical understanding of reinforcement learning approaches such as RLHF, RLAIF, or policy optimization.
- Experience leading technical teams and driving architecture decisions for large-scale AI platforms.
Technical Skills
Programming
- Python
- TypeScript (preferred)
AI & LLM Frameworks
- LangGraph
- LangChain
- Model Context Protocol (MCP)
- Function Calling
- Multi-provider LLM integrations
Retrieval & Search
- Vector databases (Qdrant, Weaviate, pgvector, Pinecone, etc.)
- Hybrid Search
- Embeddings
- Retrieval Evaluation
- Reranking
Machine Learning
- LLMs
- SLMs
- RLHF
- RLAIF
- LoRA
- PEFT
- Prefix Tuning
- Supervised Fine-Tuning (SFT)
MLOps & Deployment
- Model Serving
- Quantization
- CI/CD
- Model Versioning
- Streaming APIs (SSE/WebSockets)
- Inference Optimization
AI Governance
- Authorization
- Guardrails
- Audit Trails
- Data Isolation
- Model Provenance
Evaluation
- LLM-as-a-Judge
- Retrieval Metrics
- AI Quality Evaluation
- Custom Validation Pipelines
Preferred Qualifications
- Experience building AI solutions for enterprise or regulated industries.
- Exposure to event-driven architectures and distributed systems.
- Familiarity with enterprise search platforms and knowledge management systems.
- Experience designing production evaluation frameworks for AI applications.
- Strong understanding of scalable cloud-native AI infrastructure.
Click on Apply to know more.