Multimodal AI Engineer (voice and vision)

Axiom Global Technologies; United States; Remote; USD 150,000 - 190,000; Full time

Job summary

Axiom Global Technologies seeks a Multimodal AI Engineer to own streaming speech and real-time computer vision for identity verification and proctoring. You will work on vision-language models in production and optimize models for constrained hardware while maintaining rigorous evaluation practices.

The role requires strong PyTorch, Python, Linux, SQL skills, and autonomous work with clean pull requests in a Kubernetes-deployed service. Excellent English communication is essential.

Qualifications

• Must have streaming speech (recognition, synthesis, or both) shipped in production with latency numbers.
• At least one vision model trained or fine-tuned on data they labeled or curated themselves, evaluated against incumbents, and shipped.
• Strong PyTorch experience.
• Audio fundamentals: voice activity detection, endpointing, resampling, feature extraction.
• Vision fundamentals: face detection/embedding, gaze estimation, object detection (Ultralytics/YOLO), OpenCV.
• Vision-language models in production or strong prompting/batching knowledge for structured output.
• Model export and optimization for constrained hardware (ONNX Runtime plus TensorRT or INT8 or CPU inference).
• Rigorous evaluation: held-out sets, confusion matrices, calibrated thresholds; honest reporting when models worsen.
• Proficiency in Python, Linux, and SQL; comfortable with Kubernetes-deployed services.
• Fluent written and spoken English; autonomous work style with clean PRs.

Responsibilities

• Owns streaming speech in and out of the interview in a real-time proctoring setup.
• Develops real-time computer vision for identity verification and proctoring.
• Maintains vision-language models in production and optimizes for edge hardware.
• Ensures reproducible experiments with proper evaluation protocols.

Skills

Streaming speech

Vision model training

PyTorch

Audio fundamentals

Vision fundamentals

Vision-language models

Model export/optimization

Evaluation discipline

Python/Linux/SQL

English communication

Tools

PyTorch

OpenCV

Ultralytics/YOLO

ONNX Runtime

TensorRT

Kubernetes

SQL

Linux

Job description

Role & responsibilities

Multimodal AI Engineer (voice and vision)

Owns streaming speech in and out of the interview and the real-time computer vision for identity verification and proctoring, including vision-language models.

Must have

• Streaming speech (recognition, synthesis, or both) shipped in production under a latency budget, with real numbers for time to first audio and end-of-utterance latency.
• At least one vision model trained or fine-tuned on data they labeled or curated themselves, evaluated against the incumbent, and shipped.
• Strong PyTorch.
• Audio fundamentals: voice activity detection, endpointing, resampling, and feature extraction.
• Vision fundamentals: face detection and embedding (InsightFace or equivalent), gaze estimation, object detection (Ultralytics or YOLO), and OpenCV.
• Vision-language models in production, or strong working knowledge of prompting and batching them for structured output.
• Model export and optimization for constrained hardware: ONNX Runtime plus at least one of TensorRT, INT8 calibration, or CPU-targeted inference.
• Evaluation discipline: held-out sets, confusion matrices, calibrated thresholds, and honest reporting when a new model is worse.
• Python, Linux, and SQL at a professional level. Comfortable inside a Kubernetes-deployed service.
• Fluent written and spoken English. Works autonomously with clean, reviewable pull requests.

Nice to have

• Fine-tuning speech recognition on accented or domain audio, or training text-to-speech voices.
• Speaker diarization, active speaker detection, or audio-visual sync.
• Liveness, anti-spoofing, or deepfake and virtual-camera detection.
• 3D reconstruction, depth estimation, or feature matching.