Amir – Python, LLM, Claude LLM

Developed and launched an end-to-end political intelligence platform, incorporating advanced AI and LLM solutions to deliver bias-reduced, accurate information across legal, political, and media domains. The platform enabled users to access reliable insights through structured data pipelines and conversational AI tools, supporting decision-making with high-quality, contextual information.

• Led full AI, data, web app, and DevOps teams through Agile SDLC with CI/CD practices;
• Developed and deployed advanced LLMs for NLP tasks (generation, summarization, translation, Q&A);
• Fine-tuned LLMs on domain-specific datasets to improve accuracy for specialized use cases;
• Built full data streaming and RAG pipelines using Databricks;
• Implemented Retrieval-Augmented Generation to enhance contextual relevance in outputs;
• Utilized LangChain to build conversational AI applications and integrate LLMs with other services;
• Applied Chain-of-Thought prompting to improve reasoning in complex tasks;
• Evaluated and mitigated bias and ethical concerns in LLM applications;
• Monitored performance metrics and optimized LLM pipelines;
• Stayed current on LLM research to ensure the adoption of cutting-edge solutions.

Developed and deployed a comprehensive media misinformation detection platform, enabling international clients to identify and analyze misinformation across multiple media channels, supporting data-driven decision-making and content integrity.

• Architected SaaS Gen AI platforms with embedded security and ethical AI frameworks;
• Designed and deployed scalable data streaming and RAG pipelines using Databricks, integrating LLM-based workflows with LangChain, LangGraph, and agentic frameworks;
• Built full-featured lakehouse architectures leveraging Delta Lake and Delta Live Tables (DLT) for real-time and batch data ingestion, implementing bronze-silver-gold layering, schema evolution, and CDC with autoscaling;
• Developed end-to-end ETL/ELT pipelines using PySpark, Spark SQL, and Databricks Workflows, automating large-scale transformations, dependency management, and data quality checks;
• Managed production-grade data pipelines on Azure Data Lake Storage Gen2, working with diverse formats (CSV, Parquet, Avro, Delta) and implementing intelligent partitioning and compaction;
• Led migration of legacy systems to Databricks, establishing best practices for performance tuning, cost optimization, governance (Unity Catalog), and secure multi-tenant environments with RBAC;
• Worked extensively with Spark and Pandas DataFrames for in-memory data processing, enrichment, and LLM pre-processing, integrating UDFs and Pandas UDFs for performance-critical workloads;
• Managed Databricks compute resources, including interactive clusters, job clusters, and SQL endpoints for real-time queries and analytics dashboards;
• Integrated Databricks with Azure Data Factory, Power BI, and MLflow to deliver data and LLM pipelines in production environments;
• Created a comprehensive Power BI analytics suite for Gen AI performance monitoring;
• Championed DevOps/DataOps using Databricks Repos, Git, and Azure DevOps for infrastructure-as-code and automated deployments;
• Mentored junior engineers on Databricks best practices, Spark optimization, and LLM integration to foster an innovation-driven engineering culture.

A full-scale intelligent physical robot to assist millions of visitors with navigation, orientation, and guidance across multiple sites. The robot interactively understood various languages and dialects, validated visitor questions in real time, and provided authoritative voice responses, delivering a seamless and engaging user experience.

• Architected enterprise-scale Generative AI solutions for robotics, implementing AI safety frameworks and ethical deployment guidelines;
• Established Gen AI integration strategies and best practices for enterprise adoption;
• Designed and implemented technical building blocks for Gen AI, including RAG architectures, vector databases, and multi-agent systems;
• Developed and prototyped solutions using OpenAI, Anthropic Claude, and Microsoft Copilot;
• Built AI-driven workflows and personal assistants leveraging ASR, voice-to-text, and LangChain;
• Created strategic roadmaps for Gen AI integration with existing systems, ensuring security and governance compliance.

A real-time machine learning integration and analytics system for tracking immunization-related diseases. The system integrates data streams from multiple sources to provide full visibility and control over clinical trials and publications on emerging diseases, vaccines, and therapeutic treatments. This first-class big data and machine learning initiative enabled the client to achieve tighter monitoring and control during the COVID-19 pandemic. Canadian Health Ministry save millions by using AI to automate R&D paper scanning and analysis. During Covid, it enabled swift detection of discoveries and solutions, helping save thousands of lives.

• Led a full team of data scientists and data engineers;
• Provided end-to-end architecture: real-time ingestion, ML big data storage, processing, and visualization;
• Built a Big Data Platform with real-time data ingestion using Azure Event Hub, Databricks, and AKS;
• Implemented layers for text scraping from websites, APIs, and news outlets;
• Researched and developed PoCs for Conversational AI, chatbots, and NLP content filtering (Azure Moderator, Q&A Maker, Cognitive Services) to detect fake news and harmful text;
• Applied advanced ML techniques in Python: BERTs, deep neural nets, PyTorch;
• Implemented NLP features: sentiment analysis, fake news detection, text classification, key phrase extraction, and summarization;
• Architected a centralized data warehouse using Azure SQL;
• Designed visual analytics dashboards with Power BI;
• Processed and analyzed social media, clinical, and scientific texts (Twitter, Reddit, Clinical Trials, PubMed, ArXiv) for Covid-related insights using NLP;
• Built a knowledge mining platform with Azure Cognitive Search for ingesting and classifying scientific articles, reporting insights with Power BI;
• Transitioned multiple data pipelines (traveler and lab test data) from on-premises to the cloud;
• Architected components using open-source standards: Apache Kafka (Azure Event Hub), SQL, Python, JSON, CSV, ODBC/JDBC, GitHub;
• Established a cybersecurity plan using VNet, AAD, and Azure Security Center;
• Led SDLC projects with CI/CD MLOps using Azure ML Service and Azure DevOps;
• Developed NLP PoCs with AWS Comprehend, S3, and Redshift.

A data platform modernization initiative by implementing an on-premise visualization layer using Power BI Report Server. The project focused on enabling secure, scalable, and interactive reporting while maintaining data residency requirements, allowing stakeholders to access up-to-date insights and drive data-informed decision-making within the organization.

• Helped shaping data strategy, R&R, and measured the maturity of the data analytics capabilities;
• Met with the analytics team, and analyzed their reporting requirements and converted this into technical reporting specifications;
• Installed and configured Power BI Report Server (part of Power BI Premium, On-Premise);
• Built several dashboards, reports, visualization components, KPIs, with M language and DAX language, on Power BI Report Server;
• Authored the set of roles and responsibilities in Power BI report server platform;
• Integrated all users into data source groups allowing them access to build their reports;
• Organized workshops and demos to increase awareness and adoption of BI.

A comprehensive data platform modernization initiative for an air navigation services organization, focusing on the optimization of data ingestion pipelines, architecture, and reporting layers to enhance operational efficiency and decision-making.

• Leveraged Cloud big data strategy and roadmap, and measured data maturity;
• Defined and implemented enterprise-wide data governance strategy;
• Created architectural standards for cloud analytics and AI implementation;
• Developed strategic roadmap for AI adoption and integration;
• Led implementation of data quality and security frameworks;
• Architected Cloud (Azure) big data platform and selected its technologies;
• Represented the Product Owner Role: Managed requirements in Azure DevOps, including listing the potential requirements in backlog, classification, prioritization, traceability, validation, approval of business and architectural requirements.
• Architected Cloud big data roadmap and platform; and its technologies;

A predictive analytics solution to forecast student retention and course success, enhancing institutional planning and student support strategies. Conducted in-depth business analysis to identify key drivers impacting student outcomes, designed and implemented a predictive analytics prototype, and defined solution specifications aligned with stakeholder objectives.

• Followed CRISP-DM methodology for data science and Machine learning projects;
• Conducted numerous business analysis sessions, and I analyzed the business complex risk prediction problem;
• Researched and PoC’ed different ML and cloud ML platforms and tools, and provided the OIRP team recommendations and guidance on ML platforms and techniques;
• Built different ML classification models in Azure Machine Learning platform, and that learn the historics of students retention in the university and predict retention for new students, using different Azure ML algorithms, including: Logistic Regression, Support Vector Machine, Neural Network, Averaged Perceptron;
• Performed different imputation and interpolation (missing-data replacement) operations, including replacing with Average, default value, and Multiple Imputation Complex Equations (MICE);
• Valuated the results quality of the classification models through different classification metrics and models, including confusion matrix, AUC, Accuracy, Precision, Recall, and F1;
• Managed the lifecycle of the prototype in Azure Machine Learning, by provisioning an Azure ML web service, by versioning it, establishing different Dev / Test environments, documenting it and sharing it for consumption with my internal end users.

AI-driven payment recovery solution that fully integrates with CRMs and payment gateways to recover failed subscription and installment transactions.

• Helped establish a data strategy and roadmap, and a strategy for data solutions, governance, modernization, standardization, and adoption;
• Architected different Data hubs and data warehouse and data marts on SQL Server, and Azure DBs;
• Built different Machine Learning models in Azure ML, Python, R, and SQL, and provided the OIRP team recommendations and guidance on ML algorithms, tools and standards;
• Created a data dictionary and data catalog for the metadata and data self-service data platform;
• Architected conceptual and logical designs including data models and data flow diagrams.

Information Systems modernization for retail companies, enhancing operational efficiency and decision-making.

• Developed BI and C# with VS and SQL Server 2008-12;
• Developed BI dashboards on: SSRS, PowerPivot, Dundas, DevExpress, QlikView and PerformancePoint/SharePoint;
• Trained clients, developers, marketers, and managers on BI.
• Worked hand-in-hand with internal and external clients and stakeholders, and documented then analyzed several analytical business and technical BI requirements and converted them into technical specifications (functional and non-functional, infrastructural and solution-based);
• Produced SSAS Cubes and Data Mining models;
• Developed a recommendation system, and SQL and MDX query builders and web services with C#, SQL, WFC, DMX;
• Produced R&D projects and white papers, saving 300K$+;
• Developed and optimized ETL in SSIS, tables, triggers, SPs;
• Produced maintenance plans, permissions, replications, backups, linking servers, jobs, and tuning;
• Built entities, reports, and customizations on Dynamics CRM.