Job Information
Cognizant Associate Director- Principal Architect in Bangalore, India
Job Title: Principal Architect - LLM Agents, SLM & Multi-Agent Frameworks
Summary: We are looking for a visionary Principal Architect to lead the end‑to‑end design and delivery of AI‑powered systems centred around LLM/SLM agents and multi-agent frameworks. This role demands a blend of deep software engineering expertise and strong data science foundations – especially in applying modern ML/DL techniques, fine-tuning SLMs, working with Knowledge Graphs (KGs) in enterprise-grade graph databases, implementing RAG variants with LLMs, and architecting agent-driven solutions using modern agentic frameworks.
The ideal candidate will bring 14+ years of engineering experience , including 5+ years in AI/ML , with demonstrable experience architecting distributed systems, full-stack applications including frontend frameworks and backend technologies like Node.js and Python, and cloud-native platforms such as AWS, Azure, or GCP. In addition to strong application development skills, the candidate should bring hands-on expertise in data integration pipelines , model development, KG-driven reasoning, agentic workflows , and ML Ops .Strong problem-solving skills, excellent communication, a passion for continuous learning and mentoring junior engineers are essential.
Responsibilities:
o Lead Architectural Design:
o Define and evolve the overall architecture for LLM-powered agents and multi-agent systems that optimize agent economics over time.
o Design highly scalable, resilient microservices and distributed workflows.
o Ensure seamless integration of AI Agents with other core systems, knowledge repositories and databases (structured + unstructured).
o Drive the development of APIs and SDKs for broader ecosystem adoption.
o Model Building, SLM Development & LLM/SLM Fine‑tuning:
o Collaborate with data scientists and ML engineers to fine-tune, distil, and evaluate SLMs/LLMs and optimize SLMs/LLMs for specific tasks and domains.
o Apply techniques such as RAG, knowledge graph completion, retrieval optimization, embeddings tuning, and model compression.
o Hands-on experience with agent frameworks like MAF, Autogen, AWS Agent Framework, LangGraph etc.
o Build and maintain evaluation pipelines using tools like Datadog, LangSmith, MLFlow , or equivalent
o Stay abreast of the latest advancements in LLM research and development.
o Knowledge Graphs & Contextual Intelligence
Lead the design and integration of Knowledge Graphs, ontologies, and enterprise context graphs to enhance agent reasoning.
Work on entity resolution, relationship extraction, graph embeddings, and graph-based retrieval.
Architect KG-backed workflows to improve grounding, reduce hallucinations, and enable enterprise-aware agent behaviour.
o Prompt Engineering & LLM Integration:
o Develop and refine effective prompting strategies to maximize the performance of LLMs.
o Design and implement mechanisms for safe and reliable LLM integration.
o Address challenges related to bias, hallucinations, and other potential LLM limitations.
o ML Ops & Observability:
o Establish and maintain robust ML Ops practices, including CI/CD pipelines, model versioning, feature stores, model registries and experiment tracking.
o Implement comprehensive monitoring and observability solutions to track model performance, identify anomalies, and ensure system stability.
o Data Engineering & Pipelines:
o Architect and optimize data pipelines for ingestion, transformation, KG construction, and model training datasets.
o Ensure data governance, lineage, and high-quality semantics across DS workflows.
o Full-Stack Expertise:
o Possess deep expertise across the full stack, including:
o Frontend (optional) : React, Angular, Vue.js, or similar frameworks.
o Backend: Node.js, Python (with frameworks like Flask, Django, or FastAPI), Java, or other relevant languages.
o Database: SQL (MySQL, PostgreSQL), NoSQL (MongoDB, Cassandra), and experience with database design, optimization, and management.
o Cloud Platforms: Any 02 out of AWS, Azure, GCP (experience with serverless computing, containerization, and cloud-native technologies is a must).
o Team Leadership & Mentorship:
o Guide and mentor junior engineers in best practices for development and deployment of agents.
o Foster a culture of innovation, collaboration, and continuous learning within the team.
Qualifications:
- Proven Experience: 14+ years of experience in software engineering with a strong focus on AI/ML for at least 05 years.
· SLM/LLM Expertise: Proven experience in training/fine-tuning SLMs/LLMs (OpenAI, Claude, Llama, Mistral, etc.), including optimization, distillation, and RAG
· Knowledge Graph Skills: Experience in designing and working with KGs, graph databases, ontologies, graph embeddings, and contextual reasoning
· Architectural Strength: Ability to design large-scale distributed systems; expert in cloud-native and microservices architecture.
Data Engineering Skills: Strong experience with data modeling, pipelines, and data governance.
ML Ops & Observability: Expertise in CI/CD for ML, observability, model monitoring, and production ML workflows.
Full-Stack & Cloud Competency : Strong hands-on experience with frontend, backend, and cloud technologies.
Communication & Collaboration: Excellent communication, collaboration, and leadership skills.
Strong Problem-Solving & Analytical Skills: Ability to analyze complex problems and develop innovative solutions.
Continuous Learning: Passion for learning and staying up to date with the latest advancements in AI/ML.
Cognizant is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to sex, gender identity, sexual orientation, race, color, religion, national origin, disability, protected Veteran status, age, or any other characteristic protected by law.