Agentic AI Research Engineer

About the Role

At Zycus, we are building an enterprise-grade Agentic AI Platform capable of autonomous reasoning, planning, collaboration, memory, learning, and execution at scale. This platform will power procurement transformation for some of the world's largest organizations. We are looking for exceptional engineers and researchers who want to solve difficult AI systems problems, not just integrate APIs.

What You'll Build

You will help design and develop the core technology behind Zycus' Agentic AI Platform, including:

• Autonomous AI Agents
• Multi-Agent Collaboration Frameworks
• Agent Memory Systems
• Planning & Reasoning Engines
• Enterprise Knowledge Graphs
• Retrieval-Augmented Generation (RAG)
• Long-term Memory Architectures
• Tool-use Frameworks
• Workflow Orchestration
• Model Evaluation Platforms
• AI Safety & Guardrails
• AI Observability
• Agent Benchmarking Frameworks

What You'll Do

Research

• Read and implement cutting-edge AI research
• Reproduce state-of-the-art techniques from leading conferences and journals
• Evaluate emerging foundation models and reasoning architectures
• Prototype new ideas rapidly and validate them through experimentation

Engineering

• Convert research into production-grade software
• Build scalable AI infrastructure and services
• Design APIs, SDKs, and reusable AI platform components
• Optimize latency, cost, and reliability of AI systems

Work on problems such as

• Agent planning
• Multi-agent coordination
• Tool selection
• Context engineering
• Model routing
• Memory management
• AI evaluation
• Autonomous decision-making
• Reinforcement learning from feedback
• Long-context reasoning
• Semantic retrieval
• Knowledge representation

Experimentation

Design rigorous experiments to improve:

• Accuracy
• Reliability
• Hallucination reduction
• Agent success rates
• Cost efficiency
• Planning quality
• Benchmark performance

Requirements & Tech Stack

Technologies You May Work With

• Python
• PyTorch
• JAX
• Hugging Face
• vLLM
• Ray
• LangGraph
• DSPy
• LlamaIndex
• Kubernetes
• Docker
• Vector Databases
• Graph Databases
• MCP
• FastAPI
• GPU Computing

Experience with these tools is valuable but not mandatory. Strong fundamentals matter more than familiarity with any specific framework.

Qualifications

We are far more interested in exceptional problem-solving ability than years of experience with a specific framework. You might be a great fit if you have:

• M.Tech or PhD in Computer Science, Artificial Intelligence, Machine Learning, Robotics, Computational Linguistics, or a related discipline
• Outstanding software engineering skills in Python
• Strong foundations in algorithms, distributed systems, and software architecture
• Deep understanding of machine learning and modern AI techniques
• Curiosity to explore frontier AI research
• Ability to translate research ideas into working systems

Fresh PhDs and outstanding M.Tech graduates with exceptional research work are encouraged to apply.

Preferred Background

Experience in one or more of the following:

• Large Language Models
• Agentic AI
• Reinforcement Learning
• NLP
• Deep Learning
• Search
• Information Retrieval
• Knowledge Graphs
• Distributed Systems
• Probabilistic Reasoning
• Planning Algorithms
• Multi-Agent Systems
• Robotics AI
• Cognitive Architectures

What Makes Someone Exceptional

We would be excited if you have:

• Publications in NeurIPS, ICML, ICLR, ACL, EMNLP, CVPR, or similar conferences
• Open-source AI contributions
• Research internships
• Kaggle or AI competition achievements
• Experience building AI systems from scratch
• Strong mathematical foundations
• Curiosity to challenge existing approaches