Plata is a Mexico-founded fintech startup created by former Tinkoff executives that operates as a licensed bank.
<p>We are looking for an <strong>Applied Research Engineer</strong> for the <strong>Risks Team </strong>to create state of the art foundational deep learning models for fintech. We work with Python, Pytorch, MLFlow, AWS, Snowflake, Claude Code / OpenAI Codex / Cursor.</p> <p> </p> <p><strong>Challenges that await you:</strong></p> <ul> <li>Train self-supervised models on discrete sequences to beat the SOTA and achieve business impact in downstream tasks across PLATA, such as: credit risk, transaction anti-fraud and LTV forecasting.</li> <li> Stay on top of SotA research, applying the latest NLP and DL techniques to fintech models.</li> <li> Work with large multimodal datasets: tabular, behavioral, transactional, device and network, text, time series, graphs.</li> <li> Optimize the utilization of compute resources for both training and inference.</li> <li> Own and develop solutions end-to-end, from idea to data collection to experiments to training runs to inference optimization to evaluating impact.</li> <li> Perform rigorous evaluations.</li> <li> Write articles and speak at industry conferences.</li> </ul> <p> </p> <p><strong>What makes you a great fit:</strong></p> <ul> <li>Solid theoretical knowledge of machine learning, deep learning and NLP.</li> <li>Demonstrated experience in training and developing modern deep learning models.</li> <li> Understanding and experience with dataset design and data mixing for DL training.</li> <li>Demonstrated understanding of how to maximize the utilization of compute resources and efficiently perform experiments to validate research ideas.</li> <li> Solid communication skills.</li> <li>High autonomy and ownership.</li> <li>Ability to work even if Claude is down</li> </ul> <p> </p> <p><strong>Your bonus skills:</strong></p> <ul> <li>PhD, publications and other previous research experience.</li> <li>Experience training 1b+ transformer models. </li> <li>Experience with transformer-based architectures for non-natural language event sequences: CoLES, BERT4Rec, SASRec.</li> <li>Experience training self-supervised, embedding and ranking DL models.</li> <li>Experience in fintech and banking.</li> <li>B1 or higher English level for effective communication with an international team.</li> </ul> <div class="t396__elem tn-elem tn-elem__8559900061715530713485" data-elem-id="1715530713485" data-elem-type="text" data-field-top-value="1024" data-field-left-value="271" data-field-height-value="20" data-field-width-value="656" data-field-axisy-value="top" data-field-axisx-value="left" data-field-container-value="grid" data-field-topunits-value="px" data-field-leftunits-value="px" data-field-heightunits-value="px" data-field-widthunits-value="px" data-field-textfit-value="autoheight" data-field-top-res-320-value="1715" data-field-left-res-320-value="12" data-field-width-res-320-value="195" data-field-container-res-320-value="grid" data-field-top-res-480-value="1638" data-field-left-res-480-value="80" data-field-width-res-480-value="318" data-field-top-res-640-value="1204" data-field-left-res-640-value="65" data-field-width-res-640-value="307" data-field-top-res-960-value="1051" data-field-left-res-960-value="153" data-fields="top,left,width,height,container,axisx,axisy,widthunits,leftunits,topunits"> <div class="tn-atom"><strong><br>Our ways of working:</strong></div> </div> <div class="t396__elem tn-elem tn-elem__8559900061715530713504" data-elem-id="1715530713504" data-elem-type="text" data-field-top-value="1060" data-field-left-value="271" data-field-height-value="182" data-field-width-value="598" data-field-axisy-value="top" data-field-axisx-value="left" data-field-container-value="grid" data-field-topunits-value="px" data-field-leftunits-value="px" data-field-heightunits-value="px" data-field-widthunits-value="px" data-field-textfit-value="autoheight" data-field-top-res-320-value="1747" data-field-left-res-320-value="10" data-field-width-res-320-value="303" data-field-container-res-320-value="grid" data-field-top-res-480-value="1670" data-field-left-res-480-value="80" data-field-height-res-480-value="338" data-field-width-res-480-value="321" data-field-container-res-480-value="grid" data-field-top-res-640-value="1239" data-field-left-res-640-value="65" data-field-width-res-640-value="489" data-field-top-res-960-value="1087" data-field-left-res-960-value="150" data-fields="top,left,width,height,container,axisx,axisy,widthunits,leftunits,topunits"> <div class="tn-atom"> <ul> <li><strong>Innovative Spirit: </strong>A commitment to creativity and groundbreaking solutions</li> <li><strong>Honest Feedback:</strong> valuing open, transparent communication</li> <li><strong>Supportive Team: </strong>a strong, collaborative community</li> <li><strong>Celebrating Achievements: </strong>recognizing our wins together</li> <li><strong>High-Tech Environment:</strong> a team full of smart and revolutionary people who date to challenge the status quo of incumbent finances</li> </ul> <p> </p> </div> </div> <div class="t396__elem tn-elem tn-elem__8559900061715530713509" data-elem-id="1715530713509" data-elem-type="text" data-field-top-value="1274" data-field-left-value="271" data-field-height-value="20" data-field-width-value="656" data-field-axisy-value="top" data-field-axisx-value="left" data-field-container-value="grid" data-field-topunits-value="px" data-field-leftunits-value="px" data-field-heightunits-value="px" data-field-widthunits-value="px" data-field-textfit-value="autoheight" data-field-top-res-320-value="2110" data-field-left-res-320-value="11" data-field-width-res-320-value="210" data-field-container-res-320-value="grid" data-field-top-res-480-value="2032" data-field-left-res-480-value="80" data-field-width-res-480-value="318" data-field-top-res-640-value="1530" data-field-left-res-640-value="69" data-field-width-res-640-value="307" data-field-top-res-960-value="1300" data-f