Athar Rizvi

Senior Machine Learning Engineer at WeDo Technologies

Led the development of Ollama, a cross-platform CLI and REST API tool for managing large language models locally, containerizing microservices, and deploying quantized models using Python, FastAPI, and BentoML. Utilized AWS services including S3, SageMaker, Lambda, ECS/EKS, alongside infrastructure-as-code tools such as Terraform, Helm, and ArgoCD for multi-region deployments. Implemented RBAC with SpiceDB, observability with Prometheus and Grafana, and ensured security compliance through audit logs. Managed automated document processing and query response pipelines for Hartford using AWS Step Functions, Lambdas, LangChain, and OpenSearch vector indexing. Engineered scalable genomic and clinical ML pipelines for LabCorp utilizing TensorFlow, PyTorch, SageMaker, and Kubernetes with CI/CD automation, monitoring, and hyperparameter tuning, enhancing biomarker discovery and diagnostic predictions.

Machine Learning Engineer at Tucker-Rose Associates

Developed telehealth ML pipelines by ingesting and preprocessing logs with Python and AWS services, built patient-provider matching algorithms using scikit-learn, implemented NLP triaging via NLTK and spaCy, and time-series forecasting to optimize autoscaling policies. Engineered recommendation engines and monitored model drift in production. Built scalable ETL and fraud detection pipelines for Acorns, utilizing PySpark, XGBoost, isolation forests, and containerized microservices deployed on AWS ECS and Kubernetes. Automated CI/CD pipelines and feature store updates to support personalized savings and investment products.

Python Developer at Zoro

Developed Django-based RESTful services supporting product catalog and order management with improved performance to handle increased API traffic. Designed optimized MySQL schemas and implemented Celery tasks for bulk data processing. Integrated Elasticsearch for full-text search, improving search relevance significantly. Automated AWS infrastructure provisioning and deployments to enhance server performance and load management.

Senior Machine Learning Engineer at WeDo Technologies

Led development of Ollama, a lightweight CLI and REST API tool for local deployment of large language models, integrating Python, FastAPI, Pydantic, SQLAlchemy, and Jinja into isolated microservices. Managed optimized builds of llama_cpp and deployed quantized models, reducing inference memory usage. Orchestrated combined ML workflows using Hugging Face transformers, vector databases, and BentoML. Implemented infrastructure as code with Terraform, Docker, Kubernetes, GitHub Actions, Helm, and ArgoCD for multi-region deployments. Designed RBAC with SpiceDB and PostgreSQL, and used Prometheus and Grafana for observability. Delivered secure, low-latency inference services globally via ECS Fargate/EKS and Cloudflare CDN. Additionally, maintained robust CI/CD pipelines and reproducible build artifacts validated in ephemeral Kubernetes namespaces.

Machine Learning Engineer at Tucker-Rose Associates

Developed and maintained telehealth and financial services machine learning solutions. At Amwell, ingested and transformed telehealth session data for Amazon Redshift, built patient-provider matching algorithms with scikit-learn, and implemented NLP pipelines for patient query triage using NLTK and spaCy. Created forecasting models with ARIMA for autoscaling, collaborative filtering recommendation engines, and containerized ML models deployed on Kubernetes/EKS. Automated CI/CD pipelines using Jenkins and integrated AWS Lambda for scalable preprocessing. Monitored models with CloudWatch and developed fraud detection using isolation forests. At Acorns, built scalable ETL pipelines on AWS EMR, user segmentation frameworks, XGBoost models for saving propensity, and real-time fraud detection services, ensuring low latency inference and nightly automated retraining.

Python Developer at Zoro

Developed RESTful APIs in Django for product catalog and order management supporting significant API traffic growth. Optimized MySQL schemas and queries to reduce latency and improve page performance. Created Celery with RabbitMQ tasks for bulk CSV operations handling large volumes of data. Integrated Elasticsearch for full-text searches with custom analyzers and relevance tuning. Automated AWS provisioning and deployments using Fabric and Bash, introduced Redis caching and load balancing, halving server load.

Senior Machine Learning Engineer at WeDo Technologies

Led multiple AI and machine learning projects including development of Ollama, a cross-platform tool for managing and serving large language models locally. Managed infrastructure as code using Terraform, containerized Python microservices with Docker and Kubernetes, and orchestrated CI/CD pipelines in GitHub Actions. Leveraged AWS services including S3, SageMaker, Lambda, ECS/EKS for scalable deployments. Integrated role-based access control with SpiceDB and implemented monitoring solutions using Prometheus and Grafana. Collaborated on projects for clients such as The Hartford and LabCorp, delivering NLP pipelines for document processing, query response, predictive modeling, and real-time inference with optimized performance and observability.

Machine Learning Engineer at Tucker-Rose Associates

Worked on telehealth platform Amwell, developing data ingestion pipelines, patient-provider matching algorithms, and natural language processing workflows. Built forecasting models for session volumes and a collaborative filtering recommendation engine. Containerized ML models and services deploying on Kubernetes and AWS ECS. Automated CI/CD pipelines using Jenkins, incorporated AWS Lambda for preprocessing, and monitored production models for performance drift. Developed scalable ETL pipelines and machine learning models for Acorns, focused on user segmentation, savings propensity prediction, fraud detection, and portfolio recommendation. Ensured low-latency inference and automated retraining workflows.

Python Developer at Zoro

Developed RESTful services in Django to support product catalog, inventory, and order management with significant API traffic growth. Designed optimized MySQL schemas and complex queries reducing latency and improving page performance. Built Celery tasks with RabbitMQ for bulk data processing and integrated Elasticsearch for full-text search with custom analyzers and relevance tuning. Automated AWS provisioning and deployments with Fabric and Bash, implemented Redis caching and load balancing to improve server efficiency.

Senior Machine Learning Engineer at WeDo Technologies

Led development of Ollama, a lightweight CLI and REST API tool for downloading, inspecting, and serving large language models locally. Built isolated microservices using Python, FastAPI, and SQLAlchemy, containerized with Docker and deployed on Kubernetes. Managed ML workflows including version pinning, reproducible builds, and automated deployment using Terraform, Helm, and ArgoCD. Instrumented services with Prometheus and Grafana for monitoring and used Cloudflare CDN for low-latency global prediction caching.

Machine Learning Engineer at Tucker-Rose Associates

Developed telehealth data pipelines ingesting session logs from S3, transforming raw event streams for modeling in Amazon Redshift. Built and validated patient-provider matching algorithms with scikit-learn, implemented NLP pipelines using NLTK and spaCy for message classification, and employed ARIMA time series forecasting. Containerized ML models with Docker and deployed on Kubernetes clusters. Automated CI/CD pipelines in Jenkins, integrated AWS Lambda for lightweight pre-processing, and monitored model performance with CloudWatch metrics and custom alerts.

Python Developer at Zoro

Developed RESTful services in Django supporting product catalog, inventory, and order management with 3x api traffic growth without downtime. Designed and optimized MySQL schemas and complex join queries improving latency and page-rendering. Developed Celery + RabbitMQ tasks for bulk CSV imports/exports processing 10GB+ data. Integrated Elasticsearch for full-text search and tuned relevance scoring. Automated AWS provisioning and deployments using Fabric and Bash, adding Redis caching and load balancing to reduce server load.

Senior Machine Learning Engineer at WeDo Technologies

Led the development of Olama, a lightweight, cross-platform CLI and REST API tool to download, inspect, and serve large language models on local machines. Designed reproducible environments bundling Python and frameworks enabling isolated microservices that iterate locally before containerization. Managed optimized model builds and deployment of quantized models, integrated Huggin Face transformers with vector DB connectors and BentoML for rapid model serving in a unified environment. Automated Terraform infrastructure, Docker build pipelines, Kubernetes deployment with Helm charts, and GitOps reconciliation with ArgoCD. Instrumented Python services for monitoring with Prometheus and Grafana, enforcing role-based access controls and audit logging.

Machine Learning Engineer at Tucker-Rose Associates

Developed telehealth platform Amwell, processing and transforming telehealth session logs for modeling consumer behaviors and session forecasting. Built patient-provider matching algorithms and NLP pipelines for triaging intake messages with scikit-learn, NLTK, and spaCy. Engineered scalable forecasting models triggering auto-scaling on AWS ECS. Created a collaborative-filtering recommendation engine and containerized ML models deployed on-prem Kubernetes upgraded to EKS. Automated CI/CD pipelines in Jenkins. Integrated AWS Lambda functions for lightweight preprocessing and monitored model performance with CloudWatch and custom Python alerts.

Machine Learning Engineer at LabCorp

Designed and trained scalable TensorFlow and PyTorch pipelines on genomic and clinical datasets using AWS SageMaker notebooks and training jobs for predictive biomarker discovery. Built data ingestion workflows with AWS Glue and S3, preprocessing terabytes of lab results with PySpark. Containerized ML inference services deploying to Amazon EKS with horizontal pod autoscaling. Implemented batch transform jobs for nightly retrospectives tuning thresholds for performance drift. Orchestrated CI/CD pipelines to lint, test, build and deploy Docker images automatically. Developed real-time inference endpoints with FastAPI and gRPC deployed in Kubernetes with Istio service mesh securing traffic. Automated hyperparameter tuning and archived metadata in MLflow for reproducibility and audit compliance.

Python Developer at Zoro

Developed RESTful services in Django exposing product catalog, inventory, and order management endpoints supporting 3× API traffic growth without downtime. Designed and optimized MySQL schemas and complex JOIN queries reducing latency from 250ms to under 50ms, improving page-render performance by 40%. Built Celery + RabbitMQ tasks for bulk CSV imports/exports, price updates, and email alerts processing 10GB+ per run with retry logic. Integrated Elasticsearch for full-text search across millions of SKUs, built custom analyzers and tuned relevance scoring to boost first-page search success by 25%. Automated AWS provisioning and deployments, added Redis caching and load balancing to halve server load.

Senior Machine Learning Engineer at WeDo Technologies

Led development of Ollama, a lightweight cross-platform CLI and REST API tool for managing large language models entirely on local machines. Built reproducible environments bundling Python and key ML libraries, containerized microservices, and orchestrated deployment on AWS and Kubernetes with full CI/CD automation and monitoring. Developed a secure, scalable model serving platform integrating Hugging Face transformers, vector databases, and BentoML, optimized for inference memory and latency. Implemented robust role-based access controls and compliance auditing across multi-region Kubernetes clusters.

Machine Learning Engineer at Tucker-Rose Associates

Developed scalable machine learning pipelines for the Amwell telehealth platform, ingesting and processing session logs and clinical data to optimize patient-provider matching and forecasting. Built NLP classification pipelines for triaging patient intake messages and real-time fraud detection models using AWS Lambda and SQS. Containerized ML models and microservices, deploying them on on-prem Kubernetes and AWS EKS clusters with horizontal autoscaling and service mesh security. Automated CI/CD workflows and monitoring with Jenkins and CloudWatch, ensuring data drift detection and timely retraining.

Python Developer at Zoro

Developed RESTful Django services supporting a rapidly growing digital marketplace with optimized MySQL schemas and full-text ElasticSearch integration. Built Celery task queues for large CSV imports and email alerts with retry logic. Automated AWS provisioning and deployments via Fabric and Bash, implemented Redis caching and load balancing to reduce server load. Improved query latency and page performance significantly through schema design and query optimization.

Senior Machine Learning Engineer at WeDo Technologies

Led the development of Olla ma, a lightweight cross-platform CLI and REST API tool for managing large language models locally, incorporating Python microservices bundling with FastAPI, Pydantic, SQLAlchemy, and Jinja. Implemented optimized LLAMA_CPP model builds for efficient 4/8-bit quantization and inference memory reduction. Orchestrated model serving with Hugging Face transformers, vector DB connectors, and BentoML under unified environments. Automated deployment pipelines with Terraform, Docker, Helm, Kubernetes, and ArgoCD, ensuring reproducible builds through version pinning and validated compatibility. Integrated Prometheus and Grafana for monitoring, enabled CloudWatch alerts, and enforced RBAC with SpiceDB for metadata policy enforcement.

Machine Learning Engineer at Tucker-Rose Associates

Developed telehealth analytics for patient-provider matching using scikit-learn with Random Forest and logistic regression to optimize specialty fits and reduce average wait times. Built NLP pipelines using NLTK and spaCy to classify patient intake messages and automatically triage queries. Implemented time-series ARIMA forecasting to predict session volumes feeding autoscaling policies for Dockerized services on AWS ECS. Engineered collaborative-filtering recommendation engines with the Surprise library targeting personalized educational content. Containerized and deployed ML models and inference as microservices on Kubernetes and AWS EKS ensuring environment consistency. Automated CI/CD pipelines in Jenkins for linting, unit and integration testing, building Docker images, and pushing to Amazon ECR. Integrated AWS Lambda for lightweight text preprocessing and orchestrated fault-tolerant data flows with Amazon SQS. Monitored models and auto-triggered retraining workflows based on perf

Python Developer at Zoro

Developed RESTful services in Django supporting product catalog, inventory, and order management scaling API traffic 3× without downtime. Designed optimized MySQL schemas and complex JOINS reducing average latency from 250ms to under 50ms improving page render performance by 40%. Built Celery + RabbitMQ tasks for bulk CSV imports/exports, price updates, and email alerts processing 10GB+ per run with retry logic. Integrated Elasticsearch for full-text search across millions of SKUs, built custom analyzers for part numbers and technical specs, tuning relevance scoring to boost first-page search success by 25%. Automated AWS provisioning with Fabric and Bash adding Redis caching and load-balancing to halve server load.

Senior Machine Learning Engineer at WeDo Technologies

Led development of Ollama, a lightweight cross-platform CLI and REST API tool for managing large language models locally. Built isolated microservices environments using Python, FastAPI, Pydantic, SQLAlchemy, and Jinja, deployed with Docker and Kubernetes, and leveraged AWS services such as S3, SageMaker, Lambda, and ECS/EKS. Implemented CI/CD pipelines with GitHub Actions and ArgoCD for multi-region Kubernetes deployments with observability using Prometheus and Grafana. Optimized model deployment using quantized models and managed RBAC and metadata storage with SpiceDB and PostgreSQL.

Machine Learning Engineer at Tucker-Rose Associates

Developed scalable TensorFlow and PyTorch pipelines on genomic and clinical datasets using AWS SageMaker. Built data ingestion workflows with AWS Glue and S3, containerized ML inference services on Amazon EKS, and orchestrated CI/CD pipelines through GitHub Actions. Implemented real-time inference endpoints with FastAPI and gRPC on Kubernetes with Istio service mesh. Instrumented monitoring with Prometheus and Grafana. Automated hyperparameter tuning with SageMaker and archived experiment metadata with MLflow.

Machine Learning Engineer at Amwell (telehealth platform)

Ingested and preprocessed telehealth session logs using Python, Pandas, and NumPy. Developed patient-provider matching algorithms with scikit-learn, NLP pipelines with NLTK and spaCy, and time-series forecasting with ARIMA models. Built collaborative filtering recommendation engines using Surprise library. Containerized ML models and deployed microservices on Kubernetes and AWS ECS. Automated CI/CD pipelines with Jenkins. Integrated AWS Lambda for lightweight processing and monitored model performance with CloudWatch and custom alerts.

Python Developer at Zoro

Developed RESTful services in Django exposing product catalog, inventory, and order management endpoints. Designed and optimized MySQL schemas and complex queries, resulting in significant latency reductions and performance improvements. Built Celery and RabbitMQ tasks for bulk CSV imports/exports and price updates. Integrated Elasticsearch for full-text search with custom analyzers and relevance tuning. Automated AWS EC2, RDS, and S3 provisioning and deployments, and improved server load using Redis caching and load balancing.

Senior Machine Learning Engineer at WeDo Technologies

Led development of OLLama, a cross-platform CLI and REST API tool for downloading, inspecting, and serving large language models locally. Managed reproducible ML environments, optimized model builds with llama_cpp, and orchestrated Hugging Face transformers with rapid model serving using BentoML. Automated infrastructure as code with Terraform, container builds with GitHub Actions, and Kubernetes deployment with Helm and ArgoCD. Implemented role-based access control with SpiceDB and PostgreSQL, and monitored Python services using Prometheus and Grafana with integrated audit logs for security compliance.

Machine Learning Engineer at Tucker-Rose Associates

Developed Telehealth platform integrating primary care, mental health, and chronic condition management. Processed telehealth session logs from S3 using Python (Pandas, NumPy), transforming raw event streams into feature tables loaded into Amazon Redshift for downstream modeling. Designed and validated patient-provider matching algorithms using scikit-learn (Random Forest, logistic regression), optimizing specialty fits and minimizing wait times through A/B testing. Built NLP pipelines with NLTK and spaCy to classify patient intake messages and auto-route queries. Implemented time-series forecasting with ARIMA for session volume prediction enabling autoscaling of Dockerized services on AWS ECS. Engineered collaborative-filtering recommendation engines with Surprise library for personalized educational resources. Containerized ML models as microservices deployed to on-prem Kubernetes cluster, upgrading to EKS in 2018. Automated CI/CD pipelines in Jenkins including linting, testing, Dock

Python Developer at Zoro

Developed RESTful services in Django exposing product catalog, inventory, and order management endpoints supporting 3× growth in API traffic without downtime. Designed and optimized MySQL schemas and complex JOIN queries reducing average product lookup latency from 250 ms to under 50 ms and improving overall page-render performance by 40%. Developed Celery 3+ RabbitMQ tasks for bulk CSV imports/exports, price updates, and email alerts processing 10+ GB per run with retry logic. Integrated Elasticsearch (2.x) for full-text search across millions of SKUs, built custom analyzers for part numbers and technical specs, and tuned relevance scoring to boost first-page search success by 25%. Automated AWS (EC2, RDS, S3) provisioning and deployments via Fabric and Bash, added Redis caching and load balancing to halve server load.

Senior Machine Learning Engineer at WeDo Technologies

Led development of Olama, a lightweight, cross-platform CLI and REST API tool for downloading, inspecting, and serving large language models locally. Managed Python-based microservices using FastAPI, SQLAlchemy, and Docker, deployed on Kubernetes with CI/CD pipelines, observability, and version control. Integrated Hugging Face transformers and vector DB connectors for rapid model serving, optimized inference memory footprint with quantized models, and enforced security and audit compliance with RBAC and integrated logs.

Machine Learning Engineer at Tucker-Rose Associates

Developed automated document processing and query response system for a Fortune 500 insurance company using AWS services managed via Terraform. Designed plugin-based architecture supporting retrieval-augmented generation and summarization workflows. Implemented document ingestion triggering AWS Step Functions and Lambda for text extraction and chunk indexing into OpenSearch for sub-second retrieval. Tuned k-NN vector indices for efficient nearest neighbor lookup and integrated prompt engineering to generate context-aware answers with AWS Bedrock endpoints.

Machine Learning Engineer at LabCorp

Designed and trained scalable TensorFlow and PyTorch pipelines on genomic and clinical datasets for predictive biomarker discovery. Built data ingestion workflows with AWS Glue and PySpark to preprocess terabytes of lab results feeding feature stores. Containerized inference microservices deployed on Amazon EKS with horizontal pod autoscaling serving diagnostic predictions at low latency. Automated batch-transform jobs with SageMaker to validate model drift and recalibrate thresholds. Instrumented Prometheus and Grafana dashboards for end-to-end model monitoring.

Machine Learning Engineer at Tucker-Rose Associates

Ingested and preprocessed telehealth session logs using Python and transformed raw event streams into feature tables for modeling. Developed and validated patient-provider matching algorithms using scikit-learn optimizing specialty fits and minimizing wait times through A/B testing. Built NLP pipelines using NLTK and spaCy to classify patient intake messages and route queries automatically. Engineered collaborative filtering recommendation engines with the Surprise library to suggest educational resources. Containerized ML models and deployed to Kubernetes clusters with CI/CD in Jenkins and AWS environments.

Python Developer at Zoro

Developed RESTful services in Django exposing product catalog, inventory, and order management endpoints scaling 3x API traffic without downtime. Designed and optimized MySQL schemas and complex join queries reducing lookup latency and improving page-render performance. Built Celery + RabbitMQ tasks for bulk CSV imports/exports and price updates processing 10GB+ per run with retry logic. Integrated Elasticsearch for full-text search across millions of SKUs and custom analyzers to boost first-page search success. Automated AWS provisioning and deployment with Fabric and Bash adding Redis caching and load balancing.

Senior Machine Learning Engineer at WeDo Technologies

Led the development of Ollama, a lightweight, cross-platform CLI and REST API tool for downloading, inspecting, and serving large language models locally. Seamlessly bundled Python, FastAPI, Pydantic, SQLAlchemy, and Jinja to enable isolated microservices and iterated locally before containerizing. Managed optimized builds of llama_cpp and deployed quantized models to reduce inference memory footprint. Orchestrated model serving using Hugging Face transformers, vector DB connectors, and BentoML under a unified environment. Automated infrastructure as code with Terraform, Docker, GitHub Actions, and Helm, deploying to Kubernetes clusters with ArgoCD reconciliation. Instrumented services with Prometheus and Grafana for observability, enforcing RBAC and compliance through integrated audit logs.

Machine Learning Engineer at Tucker-Rose Associates

Ingested and preprocessed telehealth session logs from S3 using Python (Pandas, NumPy), transforming raw event streams into feature tables loaded into Amazon Redshift for downstream modeling. Developed and validated patient-provider matching algorithms with scikit-learn (Random Forest, logistic regression), optimizing specialty fits and minimizing average wait times through A/B testing. Built NLP pipelines using NLTK and spaCy to classify patient intake messages, automatically triaging queries to primary care, mental health, or chronic care workflows. Implemented time-series forecasting with ARIMA via statsmodels to predict hourly and daily session volumes, feeding autoscaling policies for Dockerized services on AWS ECS. Engineered a collaborative-filtering recommendation engine with the Surprise library to suggest follow-up educational resources based on patient history and similar user cohorts. Containerized all ML models and inference services in Docker, deploying them as microservi

Python Developer at Zoro

Developed RESTful services in Django exposing product catalog, inventory, and order-management endpoints that supported a 3× growth in API traffic without downtime. Designed and optimized MySQL schemas and complex JOIN queries, reducing average product-lookup latency from 250 ms to under 50 ms and improving overall page-render performance by 40%. Developed Celery 3+ RabbitMQ tasks for bulk CSV imports/exports, price updates, and email alerts processing 10 GB+ per run with retry logic. Integrated Elasticsearch (2.x) for full-text search across millions of SKUs, built custom analyzers for part numbers and technical specs, and tuned relevance scoring to boost “first-page” search success by 25%. Automated AWS (EC2, RDS, S3) provisioning and deploys via Fabric and Bash, added Redis caching and load-balancing to halve server load.

Senior Machine Learning Engineer at WeDo Technologies

Led development of Ollama, a lightweight cross-platform CLI and REST API tool for downloading, inspecting, and serving large language models locally. Built isolated microservices with Python, FastAPI, and SQLAlchemy, containerized with Docker, orchestrated on Kubernetes clusters with Helm and ArgoCD for multi-region deployments. Managed version pinning, reproducible builds, and integrated observability with Prometheus and Grafana. Applied AWS services including S3, SageMaker, Lambda, ECS/EKS, and Cloudflare CDN to optimize inference and authorization with RBAC. Implemented CI/CD pipelines and enhanced security and audit logging.

Machine Learning Engineer at Tucker-Rose Associates

Developed automated document processing and querying platform for a Fortune 500 insurance/investment company, supporting multiple NLP workflows including RAG and summarization. Designed plugin-based architecture, orchestrated AWS Step Functions and Lambda for data extraction and enrichment, integrated LangChain for semantic chunking and embedding generation, and managed OpenSearch domains with k-NN vector indexing. Maintained reproducible, version controlled infrastructure with Terraform. Tuned search configurations and engineered context-aware prompt pipelines with AWS Bedrock endpoints for accurate query responses.

Machine Learning Engineer at LabCorp

Designed and trained scalable TensorFlow and PyTorch pipelines on genomic and clinical datasets using AWS SageMaker notebooks and training jobs for predictive biomarker discovery. Built data ingestion workflows with AWS Glue and PySpark, containerized ML inference services deployed on Amazon EKS with autoscaling and low latency. Implemented batch transform jobs for nightly retrospectives, set up CI/CD pipelines in GitHub Actions, and used FastAPI microservices with gRPC in Kubernetes with Istio service mesh for secure communication. Instrumented monitoring dashboards in Prometheus and Grafana, automated hyperparameter tuning with SageMaker, and archived experiments in MLflow.

Machine Learning Engineer at Tucker-Rose Associates

Ingested and preprocessed telehealth session logs using Python (Pandas, NumPy), transforming raw event streams into feature tables loaded into Amazon Redshift for downstream modeling. Developed and validated patient-provider matching algorithms with scikit-learn, optimized specialty fits via A/B testing. Built NLP pipelines with NLTK and spaCy to classify patient intake messages and triage queries. Implemented time-series forecasting with ARIMA for session volumes to feed auto-scaling policies in Dockerized services on AWS ECS. Engineered collaborative filtering recommendation engines with Surprise library, containerized ML models for deployment on Kubernetes, and automated CI/CD pipelines with Jenkins and AWS Lambda orchestration.

Python Developer at Zoro

Developed RESTful services in Django exposing product catalog, inventory, and order management APIs supporting 3x growth in traffic with zero downtime. Designed and optimized MySQL schemas and complex JOIN queries, reducing latency from 250 ms to under 50 ms. Developed Celery + RabbitMQ tasks for bulk CSV imports/exports and price updates processing over 10GB per run with retry logic. Integrated Elasticsearch (2.x) for full-text search across millions of SKUs, built custom analyzers and tuned relevance scoring increasing first-page search success by 25%. Automated AWS EC2, RDS, and S3 provisioning, added Redis caching and load balancing to halve server load.

Senior Machine Learning Engineer at WeDo Technologies

Led end-to-end ML platform development on AWS (S3, SageMaker, Lambda, ECS/EKS) and Kubernetes, enabling local/offline inference with Ollama and Hugging Face transformers. Built reproducible ML pipelines from data ingestion to model deployment, containerized with Docker and orchestrated via Terraform and Helm. Implemented CI/CD with GitHub Actions, RBAC, and centralized observability using Prometheus, Grafana, and CloudWatch. Automated data ETL to S3, managed multi-region deployments, and maintained versioned artifacts for reproducibility; deployed endpoints with FastAPI and integrated retrieval via OpenSearch/Elasticsearch.

Machine Learning Engineer at Tucker-Rose Associates

Developed scalable ML workflows for Amwell telehealth platform: ingested and preprocessed telehealth session logs from S3 using Python (NumPy, Pandas); built patient-provider matching with scikit-learn; created NLP classifiers (spaCy, NLTK) to route queries to primary care/mental health workflows; implemented ARIMA-based time-series forecasting to anticipate session volumes; containerized services with Docker and deployed on AWS ECS/EKS; orchestrated CI/CD pipelines with Jenkins; used AWS Lambda for pre-processing; exposed endpoints with Prometheus dashboards for monitoring; ensured end-to-end reproducibility and monitoring.