Amit Jhariya
I design the systems, then build them — from AWS migrations that cut six-figure costs to AI agents and RAG pipelines running in production.
10+
Years shipping
$200K
Annual cost cut
6
Orgs & teams led
11+
Products shipped
About
Ten years of turning architecture decisions into shipped systems
I started as a software engineer at Bhavitha Tech in 2015, building data pipelines and marketplace integrations. Ten years and six organizations later — including a stint in London working on ARM semiconductor tooling for Sony and Canon, and a Technical Lead run at Bhavitha Tech that paid for itself with a $200K/year AWS migration — I now work as Technical Architect at Cloud Base Solutions.digital, designing AI-first enterprise platforms built around agents, RAG pipelines, and MCP integrations from day one.
That path is why I think in systems rather than tickets. The Baker Real Estate migration off Salesforce wasn't a lift-and-shift — it was a redesign of how the business's data and workflows should live on AWS. The same instinct carries into AI engineering: agents and RAG pipelines are still systems with latency budgets, failure modes, and cost curves, and I design them the same way I design everything else — most recently an audit-critical backend for the Lilly CFT project and a real-time, sensor-driven occupancy planner for workplace design at Space Matrix.
Outside of client work, I build and open-source my own tools — ElectroAi, an offline RAG desktop assistant, and GPTMail, an AI reply assistant — because the fastest way to stay sharp on a technology is to ship something real with it.
Systems-first architecture
Every feature is designed against the system it lives in — microservices boundaries, data ownership, and failure modes get decided before the first line of code.
Outcomes over output
A $200K/year AWS migration and a 40% API speedup matter more than lines shipped — impact is measured in cost, latency, and reliability.
Production-grade by default
Observability, CI/CD, and standardized code review aren't an afterthought — they ship with the first version, not the third.
Mentorship & team leverage
Leading a team means multiplying it — code reviews, coding standards, and workshops that raise the floor, not just the ceiling.
“I design the systems, then build them — from AWS migrations that cut six-figure costs to AI agents and RAG pipelines running in production.”
Skills
A stack built for shipping architecture and AI together
Experience
A decade of roles, one throughline: own the system end to end
Architecting AI-first enterprise platforms — multi-tenant, event-driven, and built around AI agents and RAG pipelines rather than bolted onto them.
- Architecting AI-first enterprise platforms and cloud-native applications on Node.js, Python, TypeScript, PostgreSQL, Docker, Kubernetes, and AWS/Azure.
- Designing scalable microservices and multi-tenant systems with event-driven architecture, focused on performance, security, and high availability.
- Leading development of AI Agents, RAG pipelines, and MCP integrations that automate enterprise workflows end to end.
- Establishing engineering standards, CI/CD pipelines, API governance, and observability practices while mentoring engineering teams and driving architectural decisions.
Featured Projects
From an offline RAG assistant to a six-figure AWS migration
11 products across AI tooling, fintech, hospitality, and enterprise real estate — each shipped as the sole engineer or as technical lead.
ElectroAi
Offline desktop RAG assistant for document chat
An open-source desktop application for fully offline LLM execution with retrieval-augmented generation over local documents (PDF, CSV). Layers in live DuckDuckGo search for real-time grounding on top of a local vector store, wrapped in a React desktop UI.
ARM IP Explorer
Semiconductor IP simulation & validation platform
Cloud-native APIs and web tooling for ARM semiconductor IP management, simulation, and orchestration, built for enterprise clients including Sony and Canon. Automated validation workflows cut testing time by 25%, backed by a dual PostgreSQL/DynamoDB architecture for performance and consistency.
GPTMail
AI email & LinkedIn reply assistant
An open-sourced browser extension that uses the OpenAI API to draft context-aware email and LinkedIn replies directly in the compose window, cutting response time by 60%.
Baker Real Estate Incorporated
Salesforce → AWS migration & document platform
Led the architecture and migration of Baker Real Estate Incorporated off Salesforce onto AWS, cutting annual operating costs by $200K while improving availability. Designed and shipped a DocuSign-integrated contract generation system that digitized signatures and compliance workflows, and lifted core API performance by 40%.
Hospitality Channel Manager
Two-way OTA inventory sync for property operators
Built the integration layer connecting a hospitality platform to major OTAs — MakeMyTrip, Goibibo, Airbnb, Booking.com — with two-way inventory sync that cut manual reconciliation effort by 50%, plus automated SIBA SEF Portugal guest-registration compliance saving 40 hours/month.
KreditAid
OCR-powered financial analytics platform
A business-insights platform integrating Tesseract OCR for automated document ingestion, cutting manual data entry by 70% for financial analysis workflows.
Criteria Corp
AI-driven talent assessment platform
Contributed to AI-driven assessment solutions for talent evaluation, optimizing candidate profiling and test analysis with generative AI on AWS Bedrock and the Vercel AI SDK.
Saltmine Workplace Platform
AI-driven real-time occupancy planning
Built backend services for workplace design and space planning, including an AI-driven occupancy system that reads live sensor data and floorplans to recommend seating and meeting-room allocation in real time.
EazyVC
Secure, real-time electronic voting platform
A modern electronic voting system built on IBM Hyperledger Blockchain, handling real-time vote data with strong security protocols end to end.
TakePrism
Restaurant order aggregator
A centralized order-management system unifying Grubhub, Uber Eats, and DoorDash into a single restaurant-facing dashboard, improving order-processing efficiency by 45%.
Lilly CFT
Audit-critical backend with full change history
Led the backend architecture for the Lilly CFT project, implementing record-level change tracking and audit trails with temporal tables so every edit to critical business data stays traceable, plus RBAC and secure API contracts for sensitive operations.
Contact
Have a system worth building? Let's talk.
Open to Technical Architect, AI Engineering, and consulting engagements — reach out directly or send a message below.
Based in Bangalore, India — open to remote and relocation-friendly roles.