Agentic AI Engineer
Position Type: Full-Tme
Company: SpadEx Inc.
Location: Remote

The Vision
Transforming artificial into native intelligence. We make enterprise AI an owned asset, not a rented liability.

The Mission
Embedding dynamic routing into every model call. We slash the token tax, boost AI accuracy, and make agentic integration effortless.

The Role
SpadEx Inc. is building the true native intelligence layer for the private enterprise agentic AI. We are eradicating the $1M+ manual MLOps tax by deploying zero-ops, per-step token usage optimization on every model call.

We are looking for an elite Agentic AI Engineer who builds like a deep-tech engineer and executes in startup-speed. You will operate strictly at the vLLM, PyTorch layer and above. You will be architecting the core intelligence, pre-training reasoning models, adapters and building the dynamic hybrid routing algorithms that allow enterprise agents to execute flawlessly.

What You Will Execute:

Foundation & Reasoning Models: Architect, pre-train, and fine-tune advanced reasoning models.



Utilize cutting-edge reinforcement learning (RLHF/RLAIF) to maximize the cognitive density of our custom models.

Agentic Architecture: Design and deploy multi-agent reasoning frameworks (CoT, ReAct, Tree of Thoughts) that operate securely within our micro-agent ecosystem.

Agile Routing Algorithms: Build and refine the algorithmic intelligence that dynamically assesses and routes workloads between local SLMs and cloud models, driving external token costs to absolute zero without sacrificing accuracy.

Zero-Friction Integration: Ensure that complex agentic reasoning capabilities can be seamlessly integrated into existing enterprise workflows.

Minimum Qualifications:

Education: B.S. in Computer Science with core project/research experience on Artificial Intelligence, Machine Learning, or a closely related quantitative field.

From-Scratch Execution: Undeniable, provable track record of architecting, pre-traini

📌 Agentic AI Engineer (India)
🏢 SpadEx
📍 India