Applied AI Engineer & Researcher at Interfaze

The Role We're building Interfaze, the first LLM purpose-built for deterministic developer tasks. As an Applied AI Engineer & Researcher, you'll own the model side of our stack: training, evaluating, and shipping the small, task-specific models that make our OpenAI-compatible API faster, cheaper, and more reliable than frontier alternatives.

What You'll Do

• Design and train hybrid architectures that pair traditional deep-learning models (CNNs, encoders) with small language models for backend tasks like OCR, scraping, translation, and structured extraction
• Build evaluation harnesses and deterministic benchmarks that prove our models out-perform generic LLMs on accuracy, latency, and consistency
• Own model lifecycle: data pipelines, fine-tuning (LoRA/QLoRA), distillation, and serving on Groq LPUs and other accelerators
• Stay ahead of the research frontier — read papers, prototype quickly, and ship what works
• Partner with our full-stack engineers to expose new model capabilities through a single, consistent developer API

What We're Looking For

• 2+ years of applied ML/AI experience (industry, research, or strong open-source/portfolio work)
• Comfort with PyTorch, HuggingFace, and the modern fine-tuning toolchain
• Strong intuition for evaluation — you know that benchmarks lie if you don't design them carefully
• Curiosity about how systems run in production: latency, batching, quantization, throughput
• Nice-to-have: experience with structured-output models, tool use, or speech/vision pipelines
• Nice-to-have: published research, open-source contributions, or kaggle/competition track record