Pouria Golshanrad

AI Engineer at Wisdomise - AI Team

Developed an AI agent to identify investment and trading opportunities by analyzing content from Telegram channels, Twitter, and Reddit. Engineered an AI trading model based on Smart Money Concept. Developed a RAG (Retrieval-Augmented Generation) system for knowledge retrieval enabling the agent to interact with customers and answer crypto market related questions. Created custom tools for the agent to dynamically display topic-relevant widgets and actionable elements in response to user queries. Built a volatility prediction model using LSTM to improve risk management in trading models. Designed a Rug-Pull detection model via blockchain transactions analysis and insider relationship graph for trench tokens. Developed an AI agent to assess the risk profile of crypto wallets based on assets, transaction history, and DeFi positions. The system compares wallet risks, provides actionable recommendations to reduce exposure, and offers guided steps for users to mitigate identified risks. Des

Data Science Team Leader at Digikala - Data Team

Led the Recommendation and Personalization team at Digikala, Iran's largest online retail company, focusing on enhancing customer experience through AI-based models. Introduced a related products feature on the product details page (PDP) using image and text embeddings of products which has gained about 85% of users clicking on the PDP. Predicted customers' next purchase time for habitual products and identified reseller customers. Designed a machine learning model to forecast product sales for upcoming promotions with MAPE of 15%, aiding in the selection of top-performing products. Optimized Digikala's warehouse bin-packing using relaxation and optimization techniques, reducing parcel size which resulted in 25% reduction in warehouse costs. Generated a predictive model for warehouse replenishment, considering upcoming events and promotions for accurate predictions with 20% of MAPE. Designed and deployed an AI-driven dynamic pricing engine for over 3000 product categories, leveraging c

Data Science Team Leader at Digikala - Data Team

Led the Recommendation and Personalization team at Digikala, Iran's largest online retail company, driving AI-based improvements to customer experience. Delivered related-products feature on PDP using image and text embeddings; predicted customers' next purchase time; designed and deployed models for sales forecasting during promotions with 15% MAPE; optimized warehouse bin-packing reducing costs by 25%; developed replenishment and demand forecasting models with ~20% MAPE; built an AI-driven dynamic pricing engine for 3000+ product categories achieving 10% MAPE and reducing seller pricing outliers; implemented anomaly detection for pricing and identified top-trending products; built a demand forecasting model for daily district-level demand with 3% MAPE.

A.I Engineer at Wisdomise - AI Team

Developed an AI agent to identify investment and trading opportunities by analyzing content from Telegram channels, Twitter, and Reddit; engineered an AI trading model based on Smart Money Concept; built a RAG system for knowledge retrieval to interact with customers on crypto market questions; created custom tools for dynamically displaying topic-relevant widgets and actionable elements; built a volatility prediction model using LSTM; designed Rug-Pull detection via blockchain transaction analysis and insider relationship graphs; developed an AI wallet risk assessment system providing actionable recommendations and guided mitigation steps; designed and implemented a scalable technical analysis system for token price time series; proficient in Django and Python; familiar with AWS/Azure and Solidity for smart contracts.

Data Science Team Leader at Digikala - Data Team

Led the Recommendation and Personalization team, delivering AI-driven features such as related products via embeddings with 85% user PDP click-through; forecasted habitual purchase times and identified reseller customers; built a demand forecasting model for promotions (MAPE 15%); optimized warehouse bin-packing reducing costs by 25%; developed replenishment forecasting with 20% MAPE; engineered an AI-driven dynamic pricing engine for 3000+ product categories (MAPE 10%); implemented automatic pricing recommendations, anomaly detection for pricing outliers, and top-trending product/queries models; produced 30-day daily demand forecasts and an AI-based PLP sorting system increasing top product clicks by 20%.

AI Engineer at Wisdomise - AI Team

Led development of an AI agent to identify investment and trading opportunities by analyzing Telegram channels, Twitter, and Reddit; built an AI trading model based on Smart Money Concept; developed a RAG system for knowledge retrieval to answer crypto market questions; created custom tools for dynamic topic-relevant widgets; built a volatility prediction model with LSTM; designed Rug-Pull detection using blockchain transactions and insider relationship graphs; developed an AI agent to assess wallet risk profiles with actionable mitigation guidance; implemented a large-scale technical analysis system for token price time series; proficient with Django, Python data science stack; familiar with AWS/Azure, Web3/DeFi, and Solidity.

Data Science Team Leader at Digikala - Data Team

Led the Recommendation and Personalization team at Digikala, Iran's largest online retail company, focusing on AI-based models to enhance customer experience. Introduced a related products feature on the product details page using image and text embeddings with ~85% user click-through on PDP. Predicted customers' next purchase time for habitual products and identified reseller customers. Designed a ML model to forecast product sales for upcoming promotions (MAPE ~15%). Optimized warehouse bin-packing to reduce parcel size and cut warehouse costs by ~25%. Generated a predictive model for replenishment considering events and promotions (MAPE ~20%). Developed an AI-driven dynamic pricing engine for 3000+ product categories (MAPE ~10%) and introduced an automatic pricing recommendation system increasing marketplace revenue by ~15%. Built anomaly detection for pricing outliers and identified top-trending products and queries. Implemented demand forecasting for daily demand up to 30 days ahe

A.I Engineer at Wisdomise - AI Team

Developed an AI agent to identify investment and trading opportunities by analyzing content from Telegram channels, Twitter, and Reddit. Engineered an AI trading model based on Smart Money Concept. Built a RAG system for knowledge retrieval to interact with customers and answer crypto market questions. Created custom tools for an AI agent to dynamically display topic-relevant widgets and actionable elements. Built a volatility prediction model using LSTM to improve risk management in trading models. Designed a Rug-Pull detection model through blockchain transactions analysis and insider relationship graphs. Developed an AI agent to assess the risk profile of crypto wallets based on assets, transaction history, and DeFi positions, providing actionable recommendations. Designed and implemented a technical analysis system on large-scale token price time series data, addressing data pipeline efficiency and reliability. Proficient in Django and key Python DS/ML libraries; familiar with AWS/

Data Scientist at Tapsi - AI Team

Engineered LSTM-based models for demand and supply prediction with MAPE of 10% at Tapsi, one of Iran's largest tech startups and an online taxi-hailing platform. Developed and implemented a dynamic pricing model at Tapsi using advanced machine-learning predictions of supply-demand imbalance, ride requests data, weather conditions, and hourly/day-of-week effects, improving revenue management and customer experience significantly. Designed a predictive model utilizing LSTM networks and external factors (weather, traffic patterns, time-of-day) to forecast real-time ride demand and supply (7% MAPE), forming the core input data for dynamic pricing strategies employed by Tapsi. Estimated time of arrival using matrix factorization on time-series traffic data with 17% MAPE, achieving comparable accuracy to industry leaders with shorter data time-length.

Data Scientist at Tapsi - AI Team

Engineered LSTM-based models for demand and supply prediction with 10% MAPE; developed a dynamic pricing model using predictions of supply-demand imbalance, ride requests data, weather conditions, and hourly/day-of-week effects; designed a predictive model utilizing LSTM networks and external factors to forecast real-time ride demand and supply (7% MAPE), forming core input for dynamic pricing strategies; estimated time of arrival using matrix factorization on time-series traffic data with 17% MAPE.

Data Scientist at Tapsi - AI Team

engineered LSTM-based models for demand/supply prediction (MAPE ~10%); developed a dynamic pricing model using predictions of supply-demand, ride data, weather, and temporal effects; built a real-time ride demand/supply forecast (MAPE ~7%); implemented ETA estimation using matrix factorization on time-series data (MAPE ~17%).