Nazar – Python, Machine learning, API

It is a cross-chain peer-to-peer reputation protocol. Utilizes page rank algorithm to score and build a reputation across any entity on the internet. Unless competitors, the cold start problem is solved by analyzing more than 50M profiles, tweet history, and connections.

• Was in charge of 100GB dataset, 50M profiles, 10K scored projects, market-leading scoring, reports sold to tier1 projects
• Was responsible for SOTA NFT, Token price forecasting (0.72 F1 score)
• Made customer segmentation, targeting, and acquisition models based on Pagerank and one-hand connections on top competitors' 200M social graph connections
• Identified and interpreted trends in complex datasets
• Was in charge of popularity, Influence score, and forecast validation, 1M+ tweets processed (Intent recognition, NER, sentiment extraction)
• Managed a team of 10+ team members
• Made 8 data sources and scraped pipelines

Betting giant in Eastern Europe.

• Was in charge of 50TB Data source, 20M Profiles, was organizing and managing structure and pipelines for feature store
• Advanced SQL feature selection and aggregation, multi-database environment
• Made client behavior analysis, forecasting, transaction scoring, LTV, affiliate marketing, antifraud detection, user multi-accounting, and recommendation
• Built an ML anti-fraud model that overperforms (approx 230%) all other models in terms of revenue generation
• Managed a team of software and data engineers for full-circle product feature implementation
• Made new product feature proposition and MVP for casino transaction scoring (approx 1.5M USD per month)
• Was responsible for Cross-department communication and data-feed creation to solve other department needs

There were around 10 different projects:

• Ericsson, B2B client support system, with argent ticket management, Finding - the best feet for the problem across 5000 engineers, Automatic Q&A system, similar case detection
• Cancer Analysis and Treatment Based on bigdata
• Conversational bots
• Document classification and problem area recognition for lawyers
• Recommendation engines

• Was leading the Data Science department with 12+ Team members
• Was leading 2-3 projects in parallel as a lead data scientist
• Created 20+ MVPs and concepts in different niches like HealthTech, FinTech, AdTech, Legal, Marketing, Sales
• Made Cancer Analysis and Treatment on R
• Was in charge of customer segmentation and LTV, cross-sale, and recommendation engine
• Made Conversational bot, Intent recognition, and Compliance Audit powered by Advanced nlp approach
• Was responsible for Time series forecasting and price prediction
• Was in charge of high-priority task classification, over 500+ engineers based on their skills and previously solved issues

There were 3 different projects:

• Online leading FMCG store in Sweden
• Football gaming portal
• Mobile C2C app

• Was Responsible for the development, estimation, and delivery
• Was in charge of Emergency support of platforms
• Created CRM integration
• Was responsible for Solution Deployment