Mirai Labs

Shards Protocol is building a decentralized ad platform for the Telegram ecosystem, bridging advertisers and publishers with the precision of hypertargeting and the transparency of web3. As a Data Scientist, you’ll architect the backbone of our ad-targeting engine, leveraging on-chain/off-chain data and machine learning to redefine user acquisition in web3.

• Data Infrastructure Design: Build systems to ingest, clean, and analyze both on-chain (wallet transactions, NFT interactions) and off-chain (Telegram user behavior) data at scale.
• Ad Targeting Models: Develop ML/AI models for:
• User segmentation (e.g., classifying gamers, DeFi users, NFT collectors).
• Predictive analytics for ad performance (CTR, conversion rates).
• Real-time bidding algorithms optimized for web3 metrics (e.g., token-holder engagement).
• Privacy-First Analytics: Implement techniques like federated learning or differential privacy to ensure decentralized data compliance.
• On-Chain Insights: Track and interpret blockchain activity (e.g., Ethereum, Solana) to identify high-value user cohorts for advertisers.
• Cross-Team Collaboration: Partner with engineers to deploy models and with product teams to translate insights into ad-platform features.
• Experimentation: Design A/B tests to refine targeting strategies and measure ROI for advertisers.

• Must-Have:
• 3+ years in data science, with experience in adtech, gaming, or web3 ecosystems.
• Expertise in Python and ML libraries (Pandas, Scikit-learn, TensorFlow/PyTorch).
• Proficiency with SQL/NoSQL databases and big data tools (Spark, Kafka).
• Familiarity with blockchain data (transactions, smart contracts) and tools (The Graph, Dune Analytics, web3.py).
• Ability to distill complex datasets into actionable insights (e.g., dashboards, reports).
• Nice-to-Have:
• Experience with programmatic advertising stacks (DSPs, SSPs).
• Knowledge of privacy-preserving ML techniques.
• Fluency in English for cross-functional communication.