Artificial Intelligence Engineer

Role Description

The Artificial Intelligence Engineer will design, build, and optimize AI models and data pipelines that power VapusData's enterprise platform. In this full-time, on-site role based in Bengaluru, the engineer will develop and implement algorithms for pattern recognition, natural language processing, and neural network-based solutions. Day-to-day responsibilities include collaborating with product and engineering teams to translate business requirements into scalable AI systems, writing clean and efficient code, and integrating models into production environments. The role also involves evaluating model performance, conducting experiments, improving accuracy and efficiency, and staying current with emerging AI technologies to enhance the platform's capabilities.

What you'll actually work on

• Design the agentic layer of our platform: planning, tool use, memory, multi-step execution, and the guardrails that make it safe to run unattended
• Build a provider-agnostic abstraction across frontier and open-weight models - routing, fallback, cost/latency trade-offs, evaluation
• Run genuine R&D: benchmark model families against our own task suites, test what actually holds up on structured document reasoning, and publish findings internally that change what we ship
• Take open-source LLMs from evaluation to production — fine-tuning, quantisation, serving, and the deployment story for customers who won't send data to a hosted API

What we're looking for

• Strong hands-on Golang - this is our primary backend language, not a nice-to-have. Services are Go, gRPC, Kubernetes, Postgres
• Real experience building agentic or multi-step LLM systems, not just API calls behind a chat window
• Working exposure to open-source / open-weight models - you've served one yourself, not just read about it
• Python is a plus, especially for the research and evaluation side.
• Proficiency in Pattern Recognition and Neural Networks, with experience in designing, training, and deploying machine learning and deep learning models.
• Familiarity with AI and ML frameworks (e.g., TensorFlow, PyTorch, scikit-learn) and experience with data processing tools and APIs.
• A bias toward measuring things. Opinions about model quality are cheap; harnesses are not

Why this is worth your time

You'll own an architectural surface, not a ticket queue. Small team, direct access to customer problems, and a domain where correctness is enforced by reality rather than by taste.