Ezekiel Jones

Senior AI Engineer

ParallelStaff

Jan 2023 - Current (3 years 5 months)

• Led the development of production AI voice and backend systems using Python, supporting high-volume inbound customer interactions and automated complaint handling. • Architected free-speech conversational pipelines, moving away from legacy rule-based and IVR systems, and utilized Hugging Face Transformers and PyTorch for production NLP models, with Scikit-learn for supporting classification and evaluation workflows. • Integrated with enterprise CRM platforms, enabling real-time caller identification, automated case creation, and intelligent routing. • Conducted the implementation of intent classification, entity extraction, and subscriber verification using NLP pipelines, and secured reliable fallback paths to human agents. • Designed dat

AI Engineer

Hyperscience

Jan 2021 - Jan 2023 (2 years 1 month)

• Architected AI-driven document and image processing pipelines in Python to enable automated receipt analysis and purchase validation at scale. • Built multimodal workflows combining OpenCV-based image preprocessing, OCR, and NLP pipelines such as Pandas, NumPy, and NLTK to validate campaign eligibility. • Implemented embedding-based semantic similarity search using SentenceTransformers to improve receipt-to-campaign matching accuracy and reduce false positives. • Engineered scalable backend services with FastAPI exposing receipt ingestion, validation, and cashback decision APIs for high-volume processing. • Created low-latency data and caching layers using MongoDB for transactional storage and Redis for deduplication and validation workfl

A

Computer Vision Developer

Aurora Innovation

Jan 2019 - Jan 2021 (2 years 1 month)

• Developed computer vision pipelines for ADAS perception using camera and LiDAR data, enabling object detection, lane marking extraction, and traffic sign recognition with Python, OpenCV, PyTorch, and TensorFlow. • Enhanced AI-assisted labeling and pre-processing workflows using classical CV techniques, improving dataset quality and reducing manual annotation effort. • Engineered scalable data processing pipelines with Pandas, NumPy, PySpark, Dask, Polars, and PyArrow to handle large-scale multi-modal sensor datasets. • Designed backend CV services and APIs using FastAPI, Flask, Kafka, Redis, PostgreSQL, and MongoDB to support perception data workflows and validation. • Deployed end-to-end CV training and evaluation pipelines on AWS using

AI Developer

MedeAnalytics

Jan 2017 - Jan 2019 (2 years 1 month)

• Built real-time backend services using FastAPI and Flask for continuous ingestion of patient vital data from IoT and wearable devices, supporting high-frequency telemetry streams. • Created streaming and alerting pipelines using Kafka, Redis, and MongoDB to detect abnormal vital patterns and trigger early alerts to support clinical decision-making workflows. • Implemented data processing workflows with Pandas, NumPy, PySpark, and Dask to normalize, aggregate, and analyze time-series health data at scale. • Applied NLP techniques using NLTK and Hugging Face Transformers for text classification and intent detection across patient messages, symptom reports, and clinician notes. • Developed low-latency clinician-facing APIs using JavaScript a

About

Detail

Timeline

Résumé