Hey DataGeeks, I'm Pavan 👋

I am a Data Explorer passionate about diving into every field where data is prominent. My journey spans the full spectrum from Market Research & Supply Chain Analytics to designing Databases & ETL Pipelines. I extend this expertise into AI, developing Machine Learning and Deep Learning models with a specific focus on BERT-based Text & Semantic Analysis.

Data Scientist | ML Engineer | Product Analytics

📍 Location : Seattle, WA, USA
📞 Mobile : +1 (929) 278-4589
✉️ Email : pavan.yellathakota.ds@gmail.com
Linkedin : https://linkedin.com/in/yellatp
GitHub :   https://github.com/yellatp

Data Scientist with 3+ years of experience developing predictive models and automated data infrastructure. Proven track record in improving search precision, designing quantitative research pipelines, and implementing data-driven solutions for marketing and product growth. Skilled in bridging the gap between data engineering and stakeholder decision-making through statistical validation, A/B testing, and interactive analytics.

Domain	Stack
Languages & Databases
AWS Cloud Data
ML Frameworks
Tools & Visualization

Potsdam, NY | Jul 2025 – Present

• Backend Architecture: Designed a 0→1 Backend Ecosystem using FastAPI and PostgreSQL, orchestrating a scalable microservices bridge between Java-based core services and Python-native ML workloads.
• Cost-Efficient Infrastructure: Deployed and managed production services on DigitalOcean VPS to optimize infrastructure overhead; implemented Docker-based containerization to ensure environment parity across R&D and production.
• Advanced Retrieval (RAG): Engineered a Multi-Model "Text-to-Query" (TTQ) engine leveraging Gemini (Vertex AI) and DeepSeek APIs to enable dynamic, prompt-driven semantic search across high-dimensional talent data.
• Search Optimization: Deployed a multi-stage retrieval pipeline utilizing pgvector for Approximate Nearest Neighbor (ANN) search and CUDA-accelerated Cross-Encoders for high-precision re-ranking (targeting 38% improvement in Precision@N).
• Domain-Aware Recommendation: Developed a sector-specific ranking system using Vectorized Embeddings; shifted logic from generic role-matching to domain-expertise alignment, improving candidate-to-company fit.
• Generative Team-Composition: Built a module that translates natural language product descriptions into granular technical requirements and specific candidate matches, bridging the gap for non-technical founders.
• System Design & MCP: Led relational schema normalization, API contract definition, and R&D into Model Context Protocol (MCP) for agentic, self-correcting database interactions.

Key Technologies Used

Potsdam, NY | Sep 2024 – Apr 2025

• Portfolio Management: Managed a $650K real-capital portfolio, delivering a 51% total return and outperforming the S&P 500 benchmark by 26% (2,600 bps).
• Alternative Data Pipeline: Built a sentiment analysis engine scraping Reddit/YouTube to validate fundamental buy signals, using BERT-based sentiment scoring to overlay quantitative signals on traditional financial metrics.
• Automation: Automated the extraction of financial statements from SEC EDGAR using Python & Vertex AI, reducing data collection time by 80% for the analyst team.
• Risk Modeling: Developed Monte Carlo simulations and risk-parity models to stress-test overweight positions and quantify potential drawdowns for high-conviction trades.

Key Technologies Used

Potsdam, NY | Oct 2023 – May 2025

• Cloud Migration: Architected a centralized data lake on AWS S3, migrating legacy records to a queryable cloud environment and reducing data retrieval latency by 40%.
• ETL Optimization: Developed PySpark ETL jobs on AWS Glue to process 1M+ cross-channel events; utilized partition pruning to optimize query costs and speed.
• Uplift Modeling: Applied uplift modeling and behavioral clustering to identify high-value fan segments, optimizing marketing spend and merchandise revenue.
• Performance Analytics: Developed backend services with FastAPI and built interactive dashboards that delivered real-time performance insights to World Championship coaches.

Key Technologies Used

Hyderabad, India | Jul 2022 – Dec 2022

• Forecasting: Developed demand forecasting models (Prophet/SARIMAX) for 1,500+ SKUs, integrating exogenous variables (holidays, promotions) to improve forecast accuracy (MAPE) by 15%.
• Reporting Automation: Designed and deployed automated KPI dashboards in Oracle Analytics Cloud (OAC), saving the procurement team 12+ hours/week of manual reporting time.
• ML Workflows: Implemented GxP-compliant ML workflows on Oracle Cloud Infrastructure (OCI) with real-time alerts, achieving 99.9% uptime for critical inventory monitoring.

Key Technologies Used

Pune, India | Sep 2021 – May 2022

• Pipeline Automation: Built automated data pipelines for Tracker and Syndicated Research projects using Python and PySpark, integrating 10M+ survey records from 30+ sources and reducing processing latency by 30%.
• Statistical Analysis: Developed sampling approaches and statistical significance testing to ensure data representativeness across Middle East and Central Africa markets.
• Consumer Insights: Supported recurring monthly/quarterly client tracking projects by developing regression models and delivering insights for 10+ FMCG and Telecom clients.

Key Technologies Used

_{Last Updated: 2026 by PAVAN YELLATHAKOTA}