Amit Joshi
Solution Architect · AWS Expert · GenAI/LLM Specialist
one.amitj@gmail.com · LinkedIn · GitHub · amitj.me · Melbourne, Australia
Summary
Professional Summary.
Solution Architect with 10+ years across cloud infrastructure, DevOps, and GenAI, specializing in taking LLM systems from POC to secure, cost-efficient production on AWS. Architected 10+ production GenAI solutions, including a HIPAA-compliant multi-tenant serverless platform, and cut operational costs while lifting performance. Led Leapfrog Technology to AWS Generative AI, Agentic AI Consulting, and DevOps Competencies, and scaled Programiz's compiler infrastructure to 150,000+ concurrent and 10M+ monthly users at half the previous cost. Hands-on with Terraform, Kubernetes, Python, and serverless, and experienced leading multi-country engineering teams.
Skills
Core Skills.
Cloud Platforms
- AWS
- GCP
- Azure
- DigitalOcean
DevOps & IaC
- Terraform
- CloudFormation
- Pulumi
- Docker
- Kubernetes
- GitHub Actions
- Azure DevOps
GenAI / LLM
- OpenAI GPT
- Anthropic Claude
- Azure AI
- AWS Bedrock
- RAG
- Prompt Engineering
AI Tools
- Azure Document Intelligence
- Amazon Textract
- Pinecone
- AWS OpenSearch VectorDB
- AWS S3 Vector
- LangChain
- LangSmith
- RAGAS
- RAGChecker
Languages
- Python
- Bash
- Go
- Node.js
Monitoring & Security
- ELK
- OpenSearch
- Vanta
- Drata
- Wazuh
- CloudFlare
- Certbot
- OWASP
Databases
- MySQL
- PostgreSQL
- DynamoDB
- MongoDB
- Redis
Infrastructure
- Linux
- Windows
- Nginx
- Apache
- HAProxy
- Envoy Proxy
- Kafka
- Jenkins
- Ansible
Compliance
- HIPAA
- SOC2
- GDPR
- PCI-DSS
Experience
Professional Experience.
-
Dec 2023 – Present
Solution Architect, DevOps
Leapfrog Technology Inc. · Seattle, Washington, United States
- Organizational Leadership & AWS Strategy: Spearheaded the technical initiative to secure AWS Generative AI Competency, AWS Agentic AI Consulting Competency, AWS DevOps Competency and APN Advanced membership.
- Healthcare GenAI Evangelism: Webcast speaker with CIO Online and AWS on scaling Generative AI in healthcare.
- Enterprise GenAI Orchestration: Architected and deployed 10+ production-level LLM and GenAI solutions with optimized RAG architectures and prompt engineering.
- Scalable Multi-Tenant Architecture: Designed a HIPAA-compliant, multi-tenant serverless ecosystem with rigorous data isolation and streamlined onboarding.
- Performance & Cost Engineering: Delivered 50% reduction in operational costs and 40% improvement in product performance through strategic model selection and serverless efficiency.
- High-Throughput Data Processing: Re-engineered document processing pipelines to scale throughput by 10× (1000+ pages vs. 100+), utilizing multicloud AI services.
- Product Modernization: Redesigned legacy EdTech products into automated, AI-driven systems via LLM integration.
Addy AI — a Leapfrog Technology Inc. venture
- Architected an LLM-based solution to automate U.S. healthcare provider back-office referral workflows
- Led a cross-functional team spanning multiple countries and disciplines
- Designed fully serverless AWS architecture enabling near real-time processing
-
Dec 2021 – Dec 2023
Principal Engineer, DevOps
Leapfrog Technology Inc. · Seattle, Washington, United States
- Engineered multi-stage deployment architectures for diverse client projects
- Scaled and stabilized the DevOps team from 5 to 25+ members
- Increased supported data sources by 5× (8 to 40+ external integrations)
- Streamlined CI/CD pipelines, increasing deployment frequency by 4× (monthly to weekly)
- Led compliance initiatives for HIPAA, SOC 2, and related regulatory standards
- Re-architected an ETL pipeline from batch-based to near real-time processing
-
Dec 2020 – Dec 2021
Lead Engineer, DevOps
- Delivered DevOps solutions ensuring high availability and minimal downtime
- Reduced deployment time by 20% for a microservices-based product
- Accelerated new service onboarding by over 90%
- Built a hybrid cloud setup with seamless in-house database integration
- Optimized an AWS-based product, reducing infrastructure costs by 30%
- Automated CI/CD pipelines for a microservices architecture
-
Sep 2022 – Mar 2025
DevOps Engineer Consultant
Programiz.com · Kathmandu, Nepal
- Modernized and re-architected decade-old deployment for high availability, reliability, and observability
- Stabilized GKE-based infrastructure for the Online Compiler, achieving near-zero downtime
- Engineered secure custom sandboxing solution for safer code execution
- Reduced infrastructure costs by 50%, scaled to 10+ compilers serving 150K+ concurrent and 10M+ monthly users
- Partnered with Security team to identify, remediate, and mitigate vulnerabilities
-
Jul 2020 – Dec 2020
Engineering Manager
Hamro Patro, Inc. · Kathmandu, Nepal
- Led the engineering team to build and launch products within the HamroPatro app
- Designed system architecture for cross-border mobile recharge and money transfer services
- Scaled the engineering team in both headcount and technical expertise
-
Jul 2017 – Jun 2020
Research Engineer / Sr. Research Engineer / Team Lead
Hamro Patro, Inc. · Kathmandu, Nepal
- Built HamroStack, an in-house deployment platform, migrating from GCP App Engine to multi-region self-managed infrastructure
- Led R&D initiatives for blockchain-based solutions across Bitcoin, Ethereum, and Hyperledger
- Implemented server monitoring and observability systems using Prometheus and Grafana
-
May 2016 – Jun 2017
Software Engineer
Javra Software Pvt. Ltd. · Kathmandu, Nepal
Speaking
Speaking Engagements & Articles.
-
GenAI for Healthcare Providers and Digital Health Companies
CIO.com webcast · co-hosted by Amazon Web Services (AWS) and Leapfrog Technology.
-
Dawn of the Agents
AWS User Group meetup · GenAI maturing into the age of autonomous agents.
-
In the Age of Vibes, Someone Has to Be the Checker
LinkedIn article · exploring vibe coding and the critical role of human reviewers when AI generates code.
Education
Education.
Bachelor of Engineering in Computer Engineering Kathmandu University · 2015
Resume
Full Resume & Contact.
References and recommendations available on request.