Building the Future of AI
I'm Anuj — architect of M4ST, building autonomous agent systems from Bareilly, India. I craft intelligent agents that perceive, reason, and act.
agent count: 4 active
uptime: 99.97%
region: bareilly.in
About M4ST
Based in Bareilly, India, I architect autonomous agent systems under the M4ST moniker — a studio focused on building AI agents that don't just respond, but act. From perception pipelines to decision layers to tool-use orchestration, every system is designed for real-world autonomy.
My work spans the full stack of agentic AI: multimodal models, recursive reasoning loops, memory architectures, and secure sandboxed execution. I believe the next paradigm isn't bigger models — it's better agents.
Agent Architecture
Multi-agent orchestration & reasoning
Full-Stack AI
Models to deployment pipelines
Secure by Design
Sandboxed tool-use & data isolation
Skills & Expertise
Autonomous Agents
Multi-agent systems, tool-use orchestration, recursive reasoning loops, memory architectures
AI / ML
LLM fine-tuning, RAG pipelines, embedding models, multimodal perception, RLHF
System Architecture
Distributed systems, microservices, event-driven architectures, message queues
Full-Stack Dev
TypeScript, React, Node.js, Python, Go, Rust, scalable API design, real-time systems
Security & Sandboxing
Container isolation, capability-based security, audit logging, prompt injection defense
Data & Infra
Vector databases, Kubernetes, CI/CD, observability, infrastructure-as-code, edge deployment
Featured Projects
M4ST Core
Autonomous multi-agent orchestration framework with dynamic tool discovery, recursive planning, and sandboxed execution environments.
View Project →Athena Memory
Long-term memory architecture for agents using hybrid vector/graph retrieval with automatic knowledge consolidation and episodic recall.
View Project →OmniSight
Multimodal perception layer enabling agents to process and reason over images, video streams, audio, and structured documents in real time.
View Project →Vault Agent
Capability-based security sandbox for agent tool execution with granular permissions, audit trails, and real-time threat monitoring.
View Project →Development Timeline
M4ST Platform Launch
Released the M4ST autonomous agent platform with multi-agent orchestration, memory systems, and sandboxed tool-use. Open-source agent SDK in active development.
Agent Systems Research
Deep research into recursive reasoning, tool-use planning, and memory-augmented LLM architectures. Built Athena Memory and OmniSight perception layer.
AI Infrastructure & Full-Stack
Designed and deployed large-scale AI infrastructure pipelines. Built real-time data processing systems and secure execution environments for AI workloads.
Foundations
Started building in Bareilly. Full-stack development, distributed systems, and the first experiments with LLM-based agents that would become M4ST.
Get in Touch
Building the future of autonomous agents from Bareilly, India. Whether you want to collaborate, invest, or deploy M4ST agents — let's talk.