Mesha – Python, Golang, FastAPI

It's an AI chatbot platform designed for internal use by the company's support for analysis, troubleshooting, and debugging of tickets, logs, and alerts.

• Led end-to-end development of a production LLM-powered support assistant for a Fortune 50 telecom client, covering architecture, implementation, deployment, and ongoing optimization.
• Designed and implemented LLM orchestration pipelines using Promptflow and LangGraph, enabling ticket classification, device diagnostics, and automated customer support workflows.
• Built and maintained a scalable AI backend with FastAPI and REST APIs, integrating OpenAI and Hugging Face models, including local model deployment for specific inference use cases.
• Architected the data layer using PostgreSQL and Elasticsearch to support structured data storage, semantic search, and high-volume log retrieval for support operations.
• Deployed and operated the system on Azure with full CI/CD automation via Azure DevOps, ensuring reliable, repeatable production releases.
• Delivered full-stack ownership, contributing across back-end services, AI pipelines, front-end visualization (React), data ingestion, and observability.
• Led architectural refactors to improve performance, reliability, and long-term maintainability of a live, business-critical system.
• Collaborated directly with stakeholders while operating with high autonomy, translating ambiguous business requirements into production-ready AI solutions.

It's a grocery delivery service.

• Led the modernization of back-end systems by migrating a large Java/Kotlin microservices ecosystem to Go, delivering significantly faster execution, ~10× lower memory usage, and substantially improved CPU utilization.
• Designed and optimized high-performance Go APIs, resulting in major reductions in response latency and a measurable improvement in end-user experience.
• Implemented a centralized DataDog logging and observability framework across 54 Go services, dramatically improving debugging efficiency and system transparency in production.
• Built a Brazilian taxation middleware service in Go, enabling compliant tax processing and seamless integration with downstream systems.
• Architected an event-driven backend using GCP Pub/Sub to support reliable, scalable processing of taxation and financial events.
• Leveraged GCP Cloud Storage and Cloud Functions to enhance system scalability, resilience, and operational efficiency in a distributed microservices environment.

It's an enterprise-grade telemedicine platform that serves patients via mobile apps and doctors via the web, supporting appointment scheduling, medical questionnaires, live consultations, and healthcare data workflows.

• Designed and maintained high-performance Go back-end services and APIs, achieving ~30% reduction in response times across patient and doctor-facing workflows.
• Built scalable AWS-based infrastructure using Lambda, S3, Cognito, SNS, SES, and SQS, increasing overall system capacity by ~40%.
• Optimized a large PostgreSQL database with minimal ORM usage, reducing query execution times by ~25% and improving reliability under load.
• Implemented back-end workflows for doctor scheduling, medical questionnaires, and real-time consultations, supporting both mobile and browser-based clients.
• Contributed to FHIR-based healthcare data flows, ETLs, and data scraping pipelines to support interoperability and downstream clinical systems.
• Collaborated with product, front-end, and infrastructure teams to design fault-tolerant AWS architectures, resulting in ~50% higher system stability and ~20% reduction in downtime.
• Operated within HIPAA-compliant constraints, ensuring secure handling of sensitive healthcare data and access control across services.

It's a cloud-native, real-time construction cost estimating platform that helps builders and contractors generate accurate cost estimates, bids, and proposals using live cost data and scalable back-end services.

• Built and optimized RESTful APIs and microservices in Go, reducing average API response times by ~20% compared to the previous implementation and improving overall system responsiveness.
• Designed and delivered back-end features with cross-functional teams, enhancing core user flows and significantly elevating end-user experience and engagement metrics.
• Utilized AWS services to improve application performance and reliability, enabling the platform to handle increased load with minimal operational overhead.
• Led development of key back-end services that contributed to a 15% increase in user engagement through optimized data retrieval and more efficient processing of construction estimates.
• Implemented containerized deployments with Docker, Kubernetes, Helm, and Tilt, ensuring scalable orchestration and reliable releases across environments.
• Worked with GraphQL and PostgreSQL to support flexible queries and efficient structured data access, facilitating responsive and extensible APIs.
• Integrated third-party systems such as Salesforce and maintained robust scripting support via Shell/Bash for operational tooling and automation.