Andrii Mendel

• • Designed and deployed a multi-tenant shipment tracking platform for a mid-size logistics company serving 8,500+ daily active users across the U.S. and Canada, built on Node.js, Express, PostgreSQL, and Redis; reduced average query response time by 43% through query optimization and read replica routing.

• • Built a HIPAA-compliant appointment and billing API for a regional telehealth provider using Python/FastAPI, PostgreSQL, and AWS RDS, integrating with Stripe for co-pay processing and Twilio for SMS appointment reminders.

• • Architected an event-driven order processing pipeline on AWS (SQS, Lambda, DynamoDB) capable of handling 150,000+ order events per day for a regional e-commerce fulfillment client; achieved 99.97% uptime over 18 months.

• • Developed and maintained RESTful and GraphQL APIs consumed by web and mobile front-ends across 20+ client projects; wrote OpenAPI specs and maintained Postman collections for each.

• • Led migration of a monolithic PHP/MySQL application to a containerized Node.js/PostgreSQL microservices architecture on AWS ECS, cutting deployment time from 2 hours to under 8 minutes via GitHub Actions CI/CD.

• • Maintained a 4.94/5.0 average rating on Upwork across 180+ completed contracts over ten years of active freelancing.

Stack: Node.js

◦ · Express

◦ · Python

◦ · FastAPI

◦ · Go

◦ · PostgreSQL

◦ · MySQL

◦ · RDS, SQS, Lambda, ECS

• • Built a WhatsApp-based AI customer support chatbot for a regional insurance brokerage using Twilio, GPT-4o, and LangChain with intent classification, dynamic human-handoff routing, and session state management via Redis; reduced inbound support ticket volume by 34% within the first two months of deployment.

• • Developed a voice AI assistant for a healthcare scheduling provider using OpenAI Whisper (STT) and TTS, integrated with Twilio Programmable Voice and a custom NestJS orchestration layer; handled 1,200+ automated appointment bookings per month with a 91% task completion rate.

• • Architected a multi-stage RAG pipeline against a proprietary product knowledge base using LangChain, Pinecone, and pgvector, enabling semantic search and Q&A; over 30GB+ of technical documentation for an enterprise manufacturing client.

• • Implemented LLM-powered document processing workflows (contract parsing, invoice extraction, summarization) using GPT-4o with structured output and function calling, integrated into existing back-end systems via REST

APIs.

• • Built and deployed a multi-agent task orchestration system using LangGraph for an operations automation client, coordinating research, drafting, and review agents with human-in-the-loop checkpoints.

• • Fine-tuned and evaluated open-source models (Llama 3, Mistral) using QLoRA on AWS SageMaker for domain-specific classification tasks; reduced inference cost by 60% versus GPT-4 equivalents with comparable accuracy on target benchmarks.

Stack: Python

◦ · LangChain

◦ · LangGraph

◦ · OpenAI API (GPT-4o, Whisper)

◦ · Pinecone

◦ · pgvector

◦ · Twilio (WhatsApp, Voice)

◦ · AWS SageMaker

◦ · Hugging Face

◦ · Redis

◦ · FastAPI

◦ · Node.js

◦ · NestJS

• • Provisioned and managed cloud infrastructure for 15+ clients using Terraform and AWS CloudFormation; standardized IaC templates reduced onboarding time for new client environments by 60%.

• • Configured and maintained Kubernetes clusters (EKS and self-managed) for two SaaS clients, implementing horizontal pod autoscaling and automated rollbacks that eliminated unplanned downtime during traffic spikes.

• • Set up centralized observability stacks (Prometheus, Grafana, Loki) with PagerDuty alerting for three production environments; reduced mean time to resolution (MTTR) for P1 incidents from 47 minutes to under 12 minutes.

• • Implemented zero-downtime blue/green and canary deployment strategies for a fintech client processing $2M+ in daily transactions, satisfying strict SLA and audit requirements.

• • Conducted security hardening reviews and remediation for five client environments, addressing IAM misconfigurations, open security groups, and unencrypted S3 bucket policies.

Stack: Kubernetes

◦ · Docker

◦ · Terraform

◦ · AWS (EKS, ECS, CloudFormation, IAM, CloudWatch)

◦ · GCP

◦ · Prometheus

◦ · Grafana

◦ · Loki

◦ · PagerDuty

◦ · GitHub Actions

◦ · Jenkins

◦ · Ansible