Armin – AI agent development, Python, LLM

Built and productionized AI-powered translation and quality-assurance tools used by professional translators, linguistic reviewers, project managers, clients, and internal operations teams. The platform automates key parts of the localization workflow, including LLM-based translation, translation-memory and knowledge-base retrieval, automated quality estimation, multilingual content validation, and LLM-as-a-judge evaluation. Developed systems operating across multiple languages and high-volume file and segment workflows, including 24 automated linguistic quality checks. These tools improved translation accuracy and consistency while significantly reducing manual review time, processing costs, and the effort required to deliver multilingual content at scale.

• AI Research & Implementation: Conducted applied research on emerging NLP and MTQE methods, transforming peer-reviewed techniques into deployable production systems tailored to translation workflows.
• Model Development & Evaluation: Designed, fine-tuned, and evaluated transformer-based models (e.g., COMET, mBERT, GPT, custom encoder-decoder architectures) to enhance translation accuracy, fluency, and context preservation across multiple languages.
• Pipeline Automation: Architected end-to-end automated pipelines streamlining translation, evaluation, and delivery, integrating translation memory, quality estimation, and linguistic review in a continuous AI-assisted loop.
• Quality Estimation (MTQE): Led the design of an AI-driven translation quality estimation system combining reference-free and reference-based evaluation methods for consistent and explainable scoring.
• R&D Leadership: Oversaw experimental design, benchmarking, and internal research documentation; provided technical direction for prototype validation and production integration.
• LLM Integration & Contextual Translation: Developed and evaluated LLM-based translation enhancement and correction modules using OpenAI and HuggingFace APIs to improve domain-specific accuracy and tone adaptation.
• Data Engineering & Retrieval: Implemented context-retrieval mechanisms using vector databases (Pinecone, ChromaDB, OpenSearch) to enhance translation consistency and maintain semantic coherence across large corpora.
• Evaluation Frameworks: Integrated LangChain, LangSmith, and LangGraph to orchestrate multi-agent pipelines, ensure traceability, and automate evaluation of model outputs across production datasets.
• Scalable API Infrastructure: Developed FastAPI-based microservices and REST endpoints supporting concurrent evaluation and model inference, deployed on Azure AI infrastructure with optimized resource management.

Enterprise AI agent platform backed by $8M in Series A funding, enabling secure and scalable autonomous AI workflows. Contributed to an enterprise agentic AI platform for building and deploying autonomous AI agents. Delivered AI-powered conversational systems and workflow automation for enterprise clients including NVIDIA and KPMG, with a focus on accuracy, hallucination reduction, scalability, and security.

• Full-Stack Development: Designed and developed end-to-end applications using FastAPI (Python) and React/TypeScript, enabling real-time interaction with autonomous AI agents.
• Hallucination Reduction & AI Reliability: Reduced LLM hallucinations through multi-agent decomposition, agent validation layers, and iterative prompt and interaction-flow refinement.
• NVIDIA – AI-Powered Sales Platform: Built a conversational AI system integrating Lyzr agents with Salesforce CRM, delivering chat interfaces, conversation management, analytics dashboards, and AI-assisted sales workflows for lead engagement and qualification.
• KPMG – AI-Driven Supplier Management: Developed an AI-powered supplier onboarding and lifecycle management platform with conversational onboarding, risk assessment, performance monitoring, and natural-language database querying using autonomous agents and real-time WebSocket interactions.

Designed the target architecture for an enterprise MLOps platform used by data scientists, ML engineers, DevOps teams, governance teams, and business stakeholders at a major Canadian financial institution. The platform standardized the full model lifecycle—from development and validation to approval, deployment, monitoring, and retraining—using Vertex AI and automated CI/CD workflows. Assessed 10+ existing ML pipelines and worked with stakeholders across ML, DevOps, architecture, governance, and business teams to identify gaps and define a secure, scalable future-state platform with stronger automation, observability, and model governance.

• Current State Analysis: Reviewed pipelines, deployment workflows, and model promotion practices to assess MLOps maturity.
• Stakeholder Engagement: Conducted interviews with data scientists, architects, and AI governance leads to align technical architecture with compliance needs.
• Architecture Design: Co-designed a scalable, secure Vertex AI–centered platform for training, validation, deployment, and monitoring of ML/LLM models.
• Governance & Compliance: Helped define a model promotion strategy, including role-based access control, audit logs, and approval workflows.
• Observability: Proposed and documented use of Prometheus + Grafana, Cloud Monitoring, and metadata logging for real-time tracking and drift detection.
• Automation & CI/CD: Recommended and partially implemented automated retraining pipelines, model versioning, and CI/CD integrations using Terraform and GitHub Actions.
• Documentation: Co-authored MLOps assessment reports, a harmonized roadmap, and playbooks for model lifecycle operations in highly regulated environments.

Built and validated an AWS-based MLOps platform designed for data scientists and researchers at a global pharmaceutical company. The platform simplified the process of taking an ML model from experimentation to production by automating data ingestion, training, hyperparameter tuning, model versioning, deployment, monitoring, and retraining. The architecture was designed to let researchers train and deploy models with significantly less infrastructure complexity and faster production workflows. A production-oriented MVP using 2–3 existing client models was implemented to validate the architecture, scalability, security, and end-to-end user experience.

• Current State Assessment: Reviewed existing data pipelines, model training flows, and deployment setups to identify scalability and automation gaps.
• Data Pipeline Optimization: Designed and built data ingestion pipelines supporting real-time and batch processing, with scalable storage and data quality monitoring for selected models.
• Model Pipeline Development: Implemented retraining workflows, hyperparameter tuning, and model versioning using SageMaker Training Jobs and Model Registry.
• Infrastructure Modernization: Architected compute infrastructure for scalable model training and deployment using AWS services such as SageMaker, RDS, and IAM.
• Automation & CI/CD Implementation: Established CI/CD pipelines for seamless deployment of data pipelines and ML models using Terraform for infrastructure as code.
• Model Deployment & Monitoring: Deployed models with real-time and batch inference, validated scalability via autoscaling endpoints, and implemented automated testing for pipeline consistency.
• Security Enhancements: Enforced security measures including RBAC, encryption for data storage, and IAM-based access controls to ensure compliance.
• Future State Architecture Design: Defined a scalable, automated ML pipeline architecture supporting continuous integration, retraining, and governance for the MVP.
• Tooling & Automation Gaps Resolution: Addressed gaps in existing CI/CD tooling to enable end-to-end automation for ML operations and data pipeline management.
• Scalability Validation: Validated architecture scalability through extensive testing of real-time inference pipelines and model performance under varied loads.

Developed three applied AI/ML products focused on automated machine learning, business-data analysis, and Canadian immigration information.

Project 1 – Feature-Gen (Academic): An open-source ML optimization library that automatically discovers useful feature transformations and feature selections to improve machine-learning models.

Project 2 – Structured-Data Reasoning Agent (Mitacs × Industry Partner): An LLM-powered analytics system that enables business analysts and non-technical users to ask questions about tabular datasets in natural language and automatically receive SQL queries, statistical or ML analyses, and visual results.

Project 3 – ImmiAI Chatbot (Mitacs × Industry Partner): A RAG-based assistant designed for people researching or applying for Canadian immigration. It retrieves information from authoritative IRCC sources and generates grounded answers based on those sources. Its outputs were also reviewed by immigration lawyers to evaluate their accuracy and usefulness.

• Algorithm R&D: Designed and coded a two-level NSGA-II genetic algorithm with bootstrapping, transformation pools, and ensemble scoring; packaged it as feature-gen and authored the publication.
• LLM Architecture & Integration: Orchestrated GPT-4–based pipelines with LangChain to parse user intent, map schema, and route queries to SQL engines, statistical tests, or XGBoost models.
• LangGraph Usage: Used LangGraph to create node layers sharing state such as dataset information and metadata.
• Prompt Engineering: Developed few-shot, ReAct, and chain-of-thought prompts; versioned them in Git and ran A/B experiments to optimize accuracy, latency, and token cost.
• Embedding & Tokenization Strategy: Selected and configured OpenAI and HuggingFace embeddings; handled context limits, token chunking, and compression to optimize semantic search accuracy and cost efficiency.
• Fine-Tuning & Adaptation: Conducted LoRA-style experiments adapting pretrained LLMs on domain-specific immigration and tabular query datasets while balancing compute overhead and response quality.
• Retrieval-Augmented Generation: Implemented a Pinecone/OpenSearch vector layer embedding scraped IRCC content and chat history with search-time reranking and citation injection.
• Vector Database Evaluation: Benchmarked OpenSearch, Pinecone, and ChromaDB for RAG pipelines based on latency, scale, multi-tenant support, and embedding compatibility.
• LLM Cost & Latency Optimization: Tuned temperature, top-p, and max tokens across prompt versions.
• Evaluation & Testing: Built automated suites measuring perplexity, ROUGE-L, BLEU, custom relevance scores, and schema consistency; logged metrics to MLflow.
• Monitoring & Observability: Instrumented FastAPI services with Prometheus exporters and Grafana dashboards for uptime, token usage, latency, and drift alerts.
• Security & Compliance: Implemented JWT authentication, IAM-scoped AWS resources, and encryption at rest and in transit to meet academic and partner data-protection requirements.
• API Design for LLM Interfaces: Exposed LLM systems via FastAPI and WebSocket APIs for real-time interaction through a Next.js frontend, supporting structured queries, file uploads, and streaming responses.

Led development of a cloud-native SaaS SEO auditing and analytics platform used by marketers, internal SEO specialists, and website owners to automatically analyze large websites and identify technical and SEO issues. The platform crawls and processes 50,000+ URLs per domain, transforms the collected data into actionable SEO and business insights, and uses LLMs to generate human-readable explanations of complex metrics and detected problems. Architected and developed 12+ interconnected microservices covering crawling, data processing, AI analysis, real-time communication, APIs, and user-facing dashboards, while improving scalability, frontend performance, and processing efficiency.

• Advanced Data Analysis & Extraction: Employed ML and NLP to distill BI from large datasets using state-of-the-art analytical techniques.
• Real-time Data Communication: Implemented WebSocket protocols to enable instantaneous data exchange among microservices, improving responsiveness and UX.
• Innovative Application Development: Utilized LangChain to develop LLM-powered applications, leveraging AI technologies.
• Performance Optimization: Deployed Redis for in-memory caching, accelerating data retrieval and processing speeds.
• Frontend Refinement: Overhauled frontend architecture to optimize performance and mitigate memory leaks, improving UI responsiveness and user satisfaction.
• Microservices Architecture: Developed and integrated 12+ microservices ensuring interoperability and resilience.
• Parallel Programming Optimization: Optimized performance using parallel programming techniques.
• Database Management: Used RDS and Firebase real-time DB, tailoring storage solutions to business requirements.
• Security Enhancement: Implemented authentication mechanisms across microservices to strengthen security.
• Legacy System Maintenance: Managed and updated a legacy Node.js project, ensuring continuity and integration with new technologies.
• Scalability & Efficiency: Leveraged Kubernetes for efficient deployment of multiple processing pods.
• Infrastructure Automation: Used Terraform for infrastructure as code to enable scalable cloud resource management.
• API Design: Crafted RESTful APIs using OpenAPI specifications for standardized data exchange.
• Cloud Solutions: Utilized AWS, Firebase, and GCP services to optimize costs while meeting operational requirements.
• CI/CD with GitHub: Used GitHub CI/CD pipelines to automate deployments and streamline workflows.

The world leader in aluminum rolling and recycling. Directed delivery of an AutoML platform that democratizes machine‑learning workflows—from data sourcing and feature engineering to automated training, validation, and one‑click deployment. The solution powers predictive‑maintenance use cases across Novelis’s global manufacturing footprint, giving non‑ML users actionable insights without coding.

• Stakeholder Engagement: Communicated design decisions and justifications to stakeholders, ensuring project alignment.
• Code Quality Assurance: Led code reviews to maintain high standards and best practices in software development.
• Backend Scalability: Designed a scalable and robust backend infrastructure to support enterprise demands.
• Design Evolution: Continuously updated and created new designs to meet evolving user needs.
• Data Visualization: Integrated Azure Power BI dashboards to surface real‑time model metrics and maintenance recommendations.
• Code Quality Leadership: Led code reviews and performance profiling (async I/O, multithreading) to sustain high standards under growing user demand.
• Node.js Feature Development: Added critical features to the platform using Node.js.
• Security Measures: Implemented authentication and authorization for data protection.
• Performance Optimization: Enhanced application performance through parallel programming.

Developed a personalized conversational AI product designed to understand users beyond the literal meaning of their messages. The system used sentiment and contextual information to recognize how a user was feeling, maintain personalized conversational context, and adapt its responses over time, creating a more natural and human-like AI companion experience. Led development across the LLM application layer, backend architecture, ML/MLOps workflows, and AWS infrastructure, while also building internal tools and deployment processes required to reliably operate and scale the AI product.

• Python and ML Development: Focused on developing Python applications with an emphasis on ML and MLOps, showcasing advanced programming skills and knowledge.
• LLM-Powered App Creation: Utilized LangChain to build sophisticated LLM-powered applications, demonstrating innovative use of AI technologies.
• AWS Infrastructure Enhancement: Strengthened the company's AWS infrastructure, ensuring robust, scalable cloud solutions.
• Code Review Excellence: Conducted thorough code reviews to uphold high standards of code quality and maintainability.
• System Design Leadership: Led the design and implementation of key company products, contributing to strategic product development.
• Technical Architecture Improvement: Played a crucial role in refining the technical architecture and enhancing development practices, aligning them with industry best practices.
• Internal Tools Development: Spearheaded the development and support of vital internal tools and platforms, improving operational efficiency.
• Chatbot Innovation: Designed and developed advanced chatbots utilizing NLP, akin to ChatGPT, to enhance user engagement and automation.
• CI/CD with GitHub Actions: Utilized for automating the deployment processes, ensuring seamless and efficient deployment workflows.

Supported a critical initiative in real-time data engineering and IoT integration, centered around ingesting and processing high-throughput streams from SCADA systems. The mission: enable predictive analytics and operational insights in the company's vision-enhancement product lines through MQTT-driven IoT pipelines, Spark-based batch processing, and cloud-native AWS tooling.

• ETL & Spark Pipelines: Developed batch and near-real-time ETL jobs using AWS Glue and PySpark to transform and enrich device telemetry; stored results in Redshift and Athena for downstream analytics.
• IoT & Microservice Coordination: Orchestrated AWS Lambda and Step Functions to connect real-time ingestion pipelines with asynchronous processing, alerting, and device state tracking.
• Monitoring & Alerting: Instrumented pipelines with CloudWatch, SQS-based dead-letter queues, and retry logic to ensure data integrity, backpressure control, and failure recovery.
• Data Quality & Validation: Enforced schema conformance and null/threshold validation across ingestion and ETL layers; aligned with analytics teams on downstream usage contracts.
• Optimization & Debugging: Resolved memory leaks in concurrent systems (async/multiprocessing) that processed live telemetry; ensured sustained uptime and throughput under load.

Contributed to a large-scale insurance modernization initiative by developing cloud-native data ingestion and processing pipelines using AWS services and Python. Collaborated with distributed teams across the UK and US to build secure, multi-tenant applications that supported real-time data workflows and analytics in the insurance domain.

• ETL Pipeline Development: Created and maintained ETL jobs using AWS Lambda and Glue, supporting ingestion and transformation of structured business data into analytical data stores.
• Batch Processing & Analytics: Designed batch-oriented data flows using AWS Step Functions and Python-based tasks to preprocess and enrich insurance-related datasets.
• Multi-Tenant Architecture: Developed modular back-end services with support for tenant-specific data isolation, caching, and scalability.
• RESTful Services & Data APIs: Built APIs in Flask and FastAPI to expose processed data for front-end consumption and analytics tools.
• Data Store Integration: Connected ingestion layers and application logic to AWS RDS, DynamoDB, and S3, optimizing data flow across environments.
• Unit & Integration Testing: Wrote Pytest-based test suites to ensure data integrity and correctness across transformations.
• CI/CD Enablement: Automated deployment of pipeline components via Jenkins and CloudFormation for consistent provisioning and delivery.

Participated in the development of Intrado Cloud Collaboration tools, aimed at enhancing meeting, messaging, and collaboration capabilities for businesses globally. This role involved optimizing and extending the functionality of RESTful APIs, SDKs, and developer documentation to support creative and productive team collaborations across distances.

• RESTful API Enhancement: Enhanced and maintained RESTful APIs, SDKs, and developer documentation, facilitating easy integration and use of collaboration tools.
• Code Review: Reviewed and provided feedback on code written by other developers, ensuring adherence to best practices.
• Carrier Integration: Developed systems for seamless connectivity with partner carriers, enhancing communication services.
• Backend Systems Architecture: Helped architect and maintain backend systems for inter-data center communication, supporting distributed infrastructure.
• Technical Documentation: Created detailed documentation and technical articles for APIs and projects, aiding developer engagement and tool utilization.
• Software Quality Assurance: Delivered high-quality software through clean coding practices, pair programming, continuous integration, testing, and code reviews.
• Frontend Development: Developed front-end features, contributing to a comprehensive user experience.
• Performance Optimization: Addressed and resolved longstanding bugs, significantly improving application performance.
• UI/UX Development: Designed highly responsive UIs with user-friendly UX, enhancing overall user satisfaction.

Led the end-to-end development of an AI-driven fashion sizing platform that recommends accurate clothing sizes based on user input. Integrated machine learning and deep learning models trained on large, domain-specific datasets, dramatically improving fit accuracy and reducing product return rates. Built scalable cloud infrastructure, rich web interfaces, and data observability systems to support and monitor the production ML lifecycle.

• ML Pipeline Development: Designed and implemented ML pipelines using Apache Airflow to automate large-scale data preprocessing, model training, evaluation, and scheduled batch inference for sizing predictions.
• Large-Scale Data Handling: Processed and cleaned large user and product datasets to support training of robust machine learning and deep learning models tailored for fashion sizing.
• ORM Integration & Schema Management: Integrated SQLAlchemy ORM and Alembic into the backend architecture, enabling efficient database access, versioned schema migrations, and long-term data model maintainability.
• Elasticsearch Integration: Implemented a custom Elasticsearch-based indexing and search engine to enable fast, fuzzy search over products and sizing recommendations.
• Observability with Kibana: Set up Kibana dashboards and log pipelines to visualize search trends, model predictions, and system health metrics across the ML backend.
• Cloud Infrastructure (AWS): Deployed scalable ML services using AWS SageMaker and S3; implemented serverless components using AWS Lambda to support dynamic business logic.
• End-to-End ML Deployment: Trained, evaluated, and deployed deep learning models into production, with full lifecycle support for versioning, rollback, and model endpoint exposure.
• Full-Stack Feature Delivery: Developed UI components in React for the size recommendation interface and built back-end Flask APIs for data submission, prediction, and feedback collection.
• Testing & Validation: Wrote automated tests for ML logic, REST endpoints, and UI interactions, ensuring accuracy, data consistency, and fault tolerance in production.

Engineered the backend for StoryWorld, a dynamic educational platform designed to immerse children in the learning of Spanish, Mandarin, and English. This project focused on creating an engaging, multilingual learning environment through interactive storytelling, leveraging technology to make language learning accessible and enjoyable for young learners.

• Backend Development: Built the backend infrastructure from the ground up using Flask, incorporating security extensions to safeguard user data.
• Database Redesign: Revamped the database schema, significantly reducing response times and enhancing the user experience.
• Query Optimization: Implemented optimization techniques to accelerate database queries, ensuring quick data retrieval.
• Database Performance: Enhanced database efficiency through sharding, distributing data across multiple machines to improve speed.
• ORM Utilization: Adopted SQLAlchemy for ORM, eliminating the need for raw SQL queries and streamlining database operations.
• Containerization: Employed Docker for the containerization of services, promoting consistency across development, testing, and production environments.