Vlad R

@v_rife

Solutions Architect

Estados Unidos
Inglés, Ruso
Parte de la información aparece en idioma inglés.
Sobre mí
I am a Machine Learning and Software Engineer with extensive experience in developing and deploying high-quality AI-powered solutions across a diverse set of industries and data types. Regardless of how you hope to leverage it - if there's data, there's a way, and I'll build it for you. • Experience developing and delivering NLP, CV, TL, RAG, QML, Time-Series, and Predictive Analytics models and Recommendation systems, among several other services • Knows Python, PyTorch, Tensorflow, Flask, Javascript, HTML, CSS, Java, Ruby, and much more. • Start-to-finish development and delivery... Lee más

Habilidades

v
v_rife
Vlad R
desconectado • 

Revisa mis servicios

Sitios web y software con IA
I will build you your dream ai tool

Porfolio

Experiencia laboral

National_Institutes of Health

Solutions Architect

National Institutes of Health • Trabajador autónomo

Apr 2025 - Present1 yr 1 mo

Contract – National Institutes of Health (NIH) •Ongoing development of agentic system to help coordinate the 27 different institutions at the NIH. •Built a Retrieval-Augmented Generation (RAG) model to generate consistently relevant, ethical, and policy-compliant responses to our clients' users' queries from a vast set of archived documents. •Spearheaded the creation of a multifaceted benchmarking pipeline for a Retrieval-Augmented Generation (RAG) model by developing and delivering a set of notebooks, designed to serve as a performance evaluation step critical to quantifying and ensuring data-driven development.

Government of the District of Columbia

2 yrs 9 mos

Machine Learning Engineer

Jan 2023 - Apr 20252 yrs 3 mos

•Built a model to predict the next most likely set of user-activity on a website (in terms of element interactions within its DOM structure) using prior such logs of user activity. •Built and published a paper regarding a variation of BERT to predict the "Big-5" personality trait scores and facets of the authors of various essays, social-media posts, and other textual data. •Built and delivered a model to predict the Schwartz value scores of the authors of various extensive multilingual textual data documents. •Built a Python script package to pinpoint and extract linguistic features from text in the Russian language for further translation, data engineering, or model training purposes. •Built and deployed a Computer Vision model to identify and optionally anonymize sensitive information after ingesting PDF, DOCX, or other textual file formats within a DoD working environment enclave (JWICS, NIPR, etc). •Conducted a thorough literature review on the most promising applications of quantum computing in the geospatial information space, selected the most viable approach, and delivered a QCNN model replicating the study’s results for further use by the client. •Built and delivered a standalone testing infrastructure to run a variety of algorithm solvers on classical and quantum chips prior to their delivery and deployment to the client. •Built and deployed an integrated pipeline infrastructure to generate a multitude of metrics and plots to benchmark the performance of classical and quantum chips while active and in use. •Built a Python script package for the multilingual machine recognition, anonymization, filtration, and translation of various textual entities of interest. •Built a speech recognition and classification model for the purpose of speedily detecting and informing agency staff of distressed on-duty personnel •Built a dashboard made to streamline internal operations, identify and address gaps in the organization’s capacity to provide services

Software Engineer

Jun 2022 - Dec 20226 mos

•Built a Computer Vision model to extract information from the clients’ equipment inspection reports relevant to the implementation of an early-warning anomaly detection system. •Built the user interface for the early-warning anomaly detection system in the form of an Elasticsearch dashboard and integrated it into the client’s internal software.