Senior ML Engineer GenAI, NLP, LLMs, Python

Amir Ali Syed

Machine Learning Engineer - Python, GenAI, LLM

Richmond, TX 703-***-**** **************@*****.*** https://www.linkedin.com/in/amiralisyed19 https://github.com/AmirAliSyed91 SUMMARY

Machine Learning Engineer with over 10 years of experience designing and deploying ML, and backend solutions across fintech, healthcare, and enterprise domains. Experienced in building NLP pipelines, multi-agent AI systems, and real-time predictive models, with a strong focus on scalable, secure, and high-performance applications. Skilled at integrating data, AI, and cloud technologies to deliver actionable insights, improve operational efficiency, and drive enterprise-ready solutions. Machine Learning Engineer Mar 2020 – Dec 2025

Hexagon Huntsville, AL

Project: https://www.adaptive-ml.com

● Developed and deployed machine learning and generative AI workflows to evaluate and serve LLMs tailored for specific enterprise use cases, leveraging Python with FastAPI, Pydantic, and SQLAlchemy for robust API and database interactions.

● Integrated llama.cpp and Hugging Face Transformers to run and fine-tune open-source LLMs, enabling efficient model experimentation, inference, and benchmarking on structured and unstructured enterprise datasets.

● Built and managed model deployment pipelines using llama-serve, facilitating reproducible, versioned model serving with optimized inference latency for production workloads.

● Architected relational and vector database solutions using PostgreSQL and VectorDBs to store embeddings, model metadata, and training datasets, enabling fast similarity search and personalized LLM outputs.

● Implemented RBAC using SpiceDB to enforce fine-grained, enterprise-level data permissions, ensuring secure multi-tenant access to models and datasets.

● Containerized services with Docker and orchestrated scalable deployments on Kubernetes (EKS/ECS), leveraging Terraform and Helm to provision infrastructure as code, enabling reproducible and highly available model hosting environments.

● Managed cloud infrastructure and storage on AWS including ECR, S3, RDS, and EFS, ensuring high-performance model storage, fast dataset access, and secure scalable compute resources.

● Set up CI/CD pipelines using GitHub Actions integrated with ArgoCD to automate model deployment, updates, and rollbacks, ensuring continuous integration and delivery of AI services with minimal downtime.

● Implemented monitoring and observability with Prometheus and Grafana, and optimized content delivery using Cloudflare CDN, providing real-time insights into model performance, API health, and end-user latency for enterprise clients.

● Built real-time streaming pipelines and WebSocket-based APIs in Python for ML model serving and multi-agent LLM systems, with GitHub repositories demonstrating production-ready implementations (pre-2023), applying principles relevant to audio/video streaming workflows.

Project: https://www.tenethealth.com

● Designed NLP pipelines to analyze unstructured patient feedback and clinical notes, helping identify trends and insights that directly contributed to improved patient satisfaction.

● Implemented a multi-agent GenAI architecture where specialized LLM agents worked together on retrieval, clinical context understanding, validation, and response generation to produce more reliable outputs.

● Improved AI chatbot accuracy by enabling agents to collaboratively retrieve and reason over medical records and treatment plans, resulting in more relevant and trustworthy responses.

● Integrated OpenAI-based LLMs into patient support systems to provide automated, real-time responses to common inquiries, increasing engagement while reducing overall response times.

● Used LangChain to coordinate multi-agent workflows and LlamaIndex to efficiently index and query healthcare data, ensuring agents had access to accurate and up-to-date information.

● Integrated VectorDB for embedding storage and similarity search, enabling personalized and context-aware patient interactions based on relevant clinical data.

● Designed the platform on AWS using EC2, RDS, and S3 to support scalability, high availability, and healthcare compliance, while meeting performance and security requirements.

● Containerized agent services with Docker and scaled them using Kubernetes to enable reliable, repeatable, and manageable microservice deployments across environments.

● Used PySpark, Pandas, and the ELK stack for large-scale data processing, monitoring, and system observability, helping teams quickly detect issues and make informed decisions. Project: https://www.greendot.com

● Developed machine learning models using scikit-learn and PySpark to analyze real-time transaction data for fraud detection, accurately identifying suspicious activity and reducing fraudulent transactions.

● Built and integrated NLP-based chatbot capabilities to handle customer support queries, improving response accuracy for common questions and reducing reliance on human agents.

● Designed rule-based and machine learning driven customer assistance workflows to handle complex financial queries, provide spending insights, and automate routine customer support tasks.

● Combined transactional data and customer profiles to deliver personalized financial insights such as budgeting guidance and spending patterns tailored to individual users.

● Implemented data pipelines to efficiently access and process structured financial data in real time, ensuring accurate and timely customer support and recommendation logic.

● Used VectorDB to store and retrieve high-dimensional embeddings for similarity search, supporting fraud detection patterns and personalized recommendation use cases.

● Leveraged AWS Lambda to run serverless functions triggered by transaction events and account updates, enabling cost-efficient scaling and near real-time notifications and alerts.

● Used Pandas and NumPy for data preprocessing, feature engineering, and numerical transformations, and built predictive models with TensorFlow to improve fraud detection and customer behavior analysis.

● Designed a microservices architecture using Docker for containerization and Kubernetes for orchestration, enabling independent scaling and reliable deployment of payment processing and transaction services.

● Integrated the ELK stack (Elasticsearch, Logstash, Kibana) for centralized logging and monitoring, providing real-time visibility into system performance and faster issue resolution across the transaction pipeline. Senior Python Developer Oct 2016 – Feb 2020

Metability LLC VA

Project: https://www.chime.com

● Utilized Python for backend development, enabling rapid development and integration of various banking features while ensuring scalability.

● Employed Django to build a secure and efficient web application, leveraging its built-in security features and ORM for managing user accounts, transactions, and financial data.

● Used DRF to build RESTful APIs, seamlessly handling account management, transactions, and user interactions.

● Used GitHub Actions to automate CI/CD pipelines, ensuring consistent and efficient build, test, and deployment processes across development and production environments.

● Chose PostgreSQL for its reliability and scalability in managing transactional data, ensuring data integrity and efficient handling of financial records.

● Developed lightweight microservices using Flask for custom data processing and integrating with third-party services within Chime’s architecture.

● Integrated Kafka for real-time data streaming and processing, essential for handling live financial transactions and delivering instant updates to users.

● Utilized NumPy and Pandas for data analysis and manipulation, generating financial reports, analytics, and insights to support decision-making and enhance user features.

● Used Elasticsearch for real-time search and analytics, Logstash for transforming and ingesting logs, and Kibana for creating interactive visualizations to monitor performance and gain insights.

● Leveraged AWS services with Lambda for serverless computing, RDS for managed PostgreSQL databases, S3 for scalable storage of financial data, Athena for querying large datasets, Docker for containerization, and Kubernetes for orchestrating containers.

Project: https://www.overstock.com

● Developed the core web application, including backend logic for user authentication, transaction processing, and data management, using Django’s built-in features and Python’s versatility.

● Built RESTful APIs to enable dynamic communication between the React front-end and Django back-end, facilitating seamless updates and interactions.

● By using Flask, I developed microservices or lightweight APIs for specific functionalities, such as integrating with third-party services or handling specialized data processing tasks.

● I used Apache Kafka to handle real-time data processing and event streaming for live inventory updates and user activity tracking, enabling real-time features.

● With PostgreSQL, I managed relational data such as user profiles, product details, and transaction records, ensuring data integrity and supporting complex queries.

● Employed Pytest for automated backend testing to ensure code quality and Selenium for end-to-end testing of the web application, verifying user interactions and functionality.

● Utilized NumPy for numerical computations and Pandas for data manipulation and analysis, generating insights from sales data and optimizing inventory management.

● Deployed the application on EC2, managed relational databases with RDS, stored static and media files in S3, and used Athena for querying large datasets, ensuring robust cloud infrastructure.

● Implemented search functionality with Elasticsearch, collected and processed log data using Logstash, and visualized metrics and logs with Kibana, providing insights and improving monitoring.

● Containerized the application components with Docker, orchestrated them with Kubernetes for scalable and efficient management, and automated CI/CD pipelines with GitHub Actions. Python Developer Jan 2015 – Sep 2016

Capital One McLean, VA

● Built and maintained backend services in Python, developing RESTful APIs using Flask to support real time financial data access and internal applications.

● Designed and optimized data pipelines using SQLAlchemy with PostgreSQL and MySQL, ensuring efficient extraction, transformation, and loading of transactional and customer metadata.

● Developed automated ETL workflows that processed large volumes of structured data using Pandas, reducing manual interventions and improving data accuracy for reporting systems.

● Deployed and managed applications on AWS, using EC2 for compute and S3 for storage, enabling scalable and reliable backend services for internal financial applications.

● Implemented unit and integration testing with pytest and unittest, and managed continuous integration and deployment with Jenkins and Git, ensuring code quality, smooth releases, and production stability. KEY COMPETENCIES

Python, Django, Flask, FastAPI, Pydantic, SQLAlchemy, DRF, TensorFlow, scikit-learn, PySpark, Pandas, NumPy, Hugging Face Transformers, llama.cpp, BentoML, LangChain, LlamaIndex, VectorDB, PostgreSQL, MySQL, SpiceDB RBAC, Docker, Kubernetes

(EKS/ECS), AWS (EC2, S3, RDS, EFS, Lambda, Athena, ECR), Terraform, Helm, GitHub Actions, ArgoCD, Prometheus, Grafana, Cloudflare CDN, ELK Stack (Elasticsearch, Logstash, Kibana), Apache Kafka, Pytest, Selenium. Education

National University of Sciences & Technology - Islamabad, Pakistan Bachelor of Science, Computer Science 2010 - 2014

Contact this candidate