Founding AI Systems Architect

Cognitive AI Systems | Neuroscience-Inspired AI | Multimodal Intelligence

At Exlords, we are building a next-generation cognitive AI computing stack inspired by neuroscience, memory systems, reasoning, and adaptive intelligence.

Our vision is to move beyond purely scale-driven AI and develop architectures capable of persistent memory, structured reasoning, multimodal understanding, and real-world autonomy.

We believe the next generation of AI will not be defined over cloud models, but by AI systems that can learn, remember, reason, adapt, and operate continuously across evolving user-centric systems with real-world automation and persistent reasoning, mirroring humans environment.

We are looking for exceptional founding AI engineers, architects, and research-focused builders who want to push the boundaries of AI innovation from India.

This is a founding-level opportunity to help shape the architecture, research direction, and implementation of a new cognitive AI stack.

What You'll Work On

Cognitive AI Research

• Memory-augmented AI systems
• Episodic, semantic, and structured memory architectures
• World models and environment representations
• Long-horizon reasoning systems
• Cognitive state management and persistent context
• Neuro-symbolic reasoning and hybrid intelligence

Model Architecture & Training

• Open-source LLMs and multimodal foundation models
• Mixture-of-Experts (MoE) systems extention
• Attention mechanism optimization and experimentation
• Cross-attention and multi-model communication
• Memory-aware reasoning architectures
• Synthetic data generation and adaptive training workflows
• Fine-tuning, evaluation, and large-scale experimentation

AI Systems Engineering

• Distributed inference and orchestration systems
• Agentic AI and autonomous execution frameworks
• Cloud training and edge deployment optimization
• Benchmarking, observability, and evaluation infrastructure
• Cognitive execution runtimes and memory pipelines

Required Technical Background

Strong experience in several of the following areas:

• Transformer architectures and internals
• PyTorch, JAX, or TensorFlow
• Large-scale model training and evaluation
• Retrieval and memory systems
• Distributed AI systems
• LLMOps and model deployment
• Multimodal AI systems
• Agentic workflows and orchestration

Preferred Research & Architecture Experience

Candidates with experience in one or more of the following areas will be strongly preferred:

• Designed or modified transformer architectures
• Implemented custom attention mechanisms
• Built memory-augmented AI systems beyond standard RAG
• Developed Mixture-of-Experts (MoE) routing systems
• Designed cross-attention fusion layers between independently trained models
• Worked on world models, planning systems, or environment representations
• Developed neuro-symbolic or hybrid reasoning architectures
• Participated in large-scale model training
• Published research papers or open-source AI architecture contributions

What We're Looking For

• Builders who enjoy working from first principles
• Engineers capable of balancing research and execution
• Individuals comfortable with ambiguity and experimentation
• People motivated by creating foundational technology rather than incremental AI applications/ wrappers.

Why Join us

Open environment and founding-level responsibility with Equity participation

Opportunity to shape a novel cognitive AI architecture

Direct involvement in research, systems design, and product direction

Access to compute infrastructure for experimentation

Work alongside with IIT Madras ecosystem startup building long-term AI infrastructure

Important: We are seeking individuals motivated by long-term ownership, deep technology creation, and meaningful equity participation. As an early-stage deep-tech startup, compensation is currently equity-focused and aligned with long-term company growth. Both part-time and full-time contributors are welcome.