Skip to content

fullstackfusions/fullstackfusions

Repository files navigation

Mihir Patel

Lead Software Engineer · AI & Data Engineering · Network Automation · Distributed Systems

Portfolio: fullstackfusions.com


About Me

I am a Lead Software Engineer with 7+ years of experience designing and delivering enterprise-scale systems across AI/ML pipelines, data engineering, backend architecture, and network automation.

At Royal Bank of Canada (RBC), I lead the development of GenAI-powered solutions within the Technology & Operations division — building RAG systems, LLM-backed compliance tools, real-time data platforms, and agentic pipelines that operate at enterprise scale.

Beyond my core role, I'm actively building:

  • FullStackFusions — a YouTube channel and technical brand for deep-dive engineering content
  • Localz — an AI-native SaaS platform for local service providers
  • Open-source tooling around agentic AI and LangGraph architectures

Core Technology Stack

Application Engineering

Programming & Frameworks

  • Tools: Python (FastAPI, Flask, Django), JavaScript/TypeScript (React, Next.js), Go (Gin, Echo, Fiber)
  • Value I Add: Clean service design, reusable architecture patterns, maintainable codebases, and faster feature delivery

APIs & Messaging

  • Tools: REST, GraphQL, gRPC, WebSockets, Webhooks
  • Value I Add: Contract-first API design, secure integrations, versioning, performance tuning, and API security with OAuth2/OIDC, JWT, RBAC/ABAC, rate limiting, and auditability

Databases

  • Tools: PostgreSQL, MongoDB, Elasticsearch, Redis, SQL Server, Cassandra
  • Value I Add: Schema and data modeling, query optimization, indexing strategy, and reliability-oriented data architecture

Data, AI & Platform

Data Engineering

  • Tools: Apache Kafka, Apache Airflow, AWS S3, AWS Glue
  • Value I Add: Resilient batch/stream pipelines, data quality guardrails, and production-ready analytics foundations

GenAI Features

  • Tools: Agentic RAG, fine-tuning, function/tool calling, prompt chaining, orchestration, and re-ranking
  • Value I Add: Business-ready AI workflows that improve accuracy, automate operations, and reduce manual effort

Platform & Environment Span

  • Tools: On-prem, cloud-native, hybrid-cloud, multi-cloud, containerized workloads, Kubernetes/OpenShift
  • Value I Add: Platform standardization, secure-by-default architecture, and smooth modernization across mixed environments

Operations & Delivery

DevOps & Cloud

  • Tools: AWS (ECS, EKS, EC2, Lambda, S3), Docker, Kubernetes, GitHub Actions, Jenkins, Terraform
  • Value I Add: Faster release cycles, stable CI/CD, secure delivery controls, and cost-aware cloud operations

Network & Observability

  • Tools: Grafana, Dynatrace, NetBrain, ExtraHop, Corvil, Elasticsearch
  • Value I Add: End-to-end visibility, faster incident triage, proactive performance tuning, and measurable reliability improvement

Leadership & Delivery

  • Tools: Architecture reviews, mentoring, shell scripting, CI/CD, delivery automation
  • Value I Add: Clear technical direction, stronger engineering practices, and higher team execution velocity

Featured Work

Paginated API Matching Pipeline Cross-dataset record matching between NetBrain and Hasura — built with retry logic, paginated API orchestration, and Airflow DAG scheduling. Focused on production-grade reliability and cross-system data integrity.

Enterprise Data Platform (MVP) End-to-end ETL pipelines extracting from Grafana, Postgres, ElasticSearch, and Dynatrace — streaming via Kafka, loading into AWS S3 data lakes — orchestrated with Airflow and GitHub Actions. Designed as a cohesive platform, not a collection of isolated pipelines.

LLM Chatbot Pipeline LangGraph-powered multi-agent chatbot with message_id tracking, real-time status streaming via FastAPI, and a custom MessageRenderer frontend supporting dynamic response types (TEXT, TEMPLATE, FORM). Built for structured state management and a responsive streaming UX.

Investigative RAG v2 & Compliance RAG v2 LangGraph + LangChain RAG systems with hybrid retrieval (Qdrant + Postgres FTS), Cohere reranking, and evidence-first response design — purpose-built for compliance and network domain investigation at RBC (12 CFR Part 30 alignment). Designed for auditability, retrieval precision, and regulatory defensibility.

Localz (In Progress) AI-native SaaS platform for local service providers — featuring AI chat, seller/customer portfolios, analytics, and payment integration. Built on FastAPI + Pydantic with a monolith-first, microservices-ready architecture.


Live Portfolio

fullstackfusions.com


Run Locally

git clone https://github.com/fullstackfusions/fullstackfusions.git
cd fullstackfusions
open index.html    # Mac
start index.html   # Windows

Blog

Blog posts are generated from _posts/*.md. After adding or removing a post, rebuild:

node scripts/build-blog.js

Preview locally:

node scripts/build-blog.js && python3 -m http.server 8080
# http://localhost:8080/blog/

See BLOG_SETUP.md for the full authoring reference.


Connect

Platform Link
📧 Email [email protected]
💼 LinkedIn linkedin.com/in/fullstackfusions
🎥 YouTube youtube.com/@fullstackfusions
✍️ Substack fullstackfusions.substack.com

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors