Artificial Intelligence Scientist

Company Description

Cyble (YC W21) is a Series B-funded global cyber intelligence start-up supported by notable VC firms and Y Combinator. We provide customers with AI-powered actionable threat intelligence to manage cyber risks effectively. Our team specializes in delivering cutting-edge solutions to help organizations and governments stay ahead of cyber threats.

We're building the next generation of memory-augmented AI systems—agents that learn, recall, and reason like humans. Our mission is to develop long-context AI that's persistent, intelligent, and adaptive. If you're passionate about solving fundamental problems in memory for AI, we want to hear from you. At Cyble, you won't just work with toy datasets or academic-scale models—you'll be applying your ideas to petabyte-scale real-world data, spanning global threat intelligence, telemetry, signals, and behavioral metadata.

This is your chance to build AI memory systems at planetary scale, work on cutting-edge LLM and neural memory infrastructure, alongside a global team of top researchers and engineers, with direct CEO-level sponsorship. Be at the forefront of semantic memory, vector embeddings, and lifelong AI learning. As this function reports to the CEO and interfaces with our U.S.-based executive team, periodic travel to Cupertino is expected. For high-impact contributors, relocation support may be offered to align with our long-term strategic planning efforts.

Key Responsibilities

• Architect and build semantic and neural memory systems for large language models (LLMs) and autonomous agents.
• Design memory-aware retrieval systems using vector databases (FAISS, Milvus, LanceDB) and embedding models (OpenAI, Hugging Face, Cohere).
• Research and implement techniques for episodic, declarative, and procedural memory inspired by cognitive and neuroscience principles.
• Integrate long-term memory modules with LLM stacks (GPT, LLaMA, Claude, Mistral, etc.).
• Contribute to both open-source frameworks and proprietary IP in RAG pipelines, agent memory, and knowledge grounding.
• Collaborate with top-tier researchers, security analysts, and the CEO himself to shape Cyble's strategic AI future.
• Collaborate with product teams to translate research into real-world AI products used globally.

Qualifications

• Bachelor's/Master's/PhD in CS, AI, Cognitive Science, Applied Math, or related disciplines (IITs, IISc, IIITs, BITS, top NITs preferred but not mandatory).
• Strong experience with PyTorch, TensorFlow, or JAX for building deep learning systems.
• Hands-on with semantic search, vector similarity, and embedding space reasoning.
• Solid grasp of LLMs and memory-augmented architectures (Transformers + Memory, RAG, kNN-augmented models, etc.).
• Research mindset—demonstrated by papers, GitHub, patents, or strong experimentation projects.
• Worked on LangChain, LlamaIndex, AutoGPT, or other memory-capable agent platforms.
• Background in neuroscience-inspired AI, knowledge graphs, or cognitive modeling.
• Strong publications or open-source visibility in AI memory systems.

Send your resume, GitHub or research portfolio, and a brief note on your work in semantic or neural memory to:

Location: Remote (India / Asia)

Function: CEO – Innovations Team