Sushant Kadam
I build backend systems and AI products that ship, with 5+ years of building production systems at every scale.
That has meant enterprise platforms at a large IT services company and order systems handling 100k+ orders a day at a high-growth startup.
These days I build AI products from scratch at an early-stage startup.
Experience
-
Senior Software Engineer at
EDRA Labs Building AI products from zero to production at an early-stage startup, owning backend architecture and agentic workflows end to end.
- Moved an AI sales copilot for Salesforce (100+ enterprise users) off Chrome-extension logic onto a FastAPI backend I set up from scratch, with OpenTelemetry tracing, LangGraph/LiteLLM for model calls, and Postgres underneath.
- Brought in Temporal for durable workflow state and Zero to stream real-time updates to the frontend from a read-only workspace table.
- Built the copilot's Ask Me Anything agent: reps ask about accounts, deals, and meetings in plain language and get answers with citations back to Salesforce.
- Cut Havi.ai page loads by 90% and built its agentic workflows for conversational slide editing, PPTX/PDF export, and shared access.
- Shipped Make A Viz, an AI diagram tool, end to end in React and FastAPI, then launched it as a ChatGPT app with SigNoz monitoring.
- Built async pipelines that process 10k+ Reddit posts a day into insights for strategy teams.
- React
- FastAPI
- Python
- LangChain
- LangGraph
- Temporal
- PostgreSQL
-
Software Development Engineer I at
Fynd Built backend services for the Order Management System powering commerce workflows across JioMart and Fynd.
- Built and scaled order workflows handling 100k+ daily orders and 200k+ during peak sale events.
- Designed return quality-check workflows and APIs that reduced fraudulent returns by 20%.
- Built asynchronous integrations across order, inventory, logistics, and fulfillment systems.
- Shipped Buy Online, Pick Up In Store workflows for international commerce markets.
- Python
- Microservices
- Kafka
- MongoDB
- Redis
- E-commerce
-
Senior Product Engineer at
LTIMindtree Built backend services for Aspect, an Intelligent Document Processing platform for extracting structured data from complex documents.
- Integrated Google Vision, AWS Textract, Kofax, and Tesseract into unified document-processing workflows.
- Led development of tabular data extraction algorithms, improving extraction accuracy by 25% and reducing false positives by 30%.
- Researched and implemented text-zoning techniques to improve document structure detection and downstream extraction.
- Python
- OCR
- AWS Textract
- Google Vision
- Document AI
- Microservices
Projects
Tech Stack
- Python
- JavaScript
- HTML
- CSS
- React
- Vite
- Next.js
- FastAPI
- Flask
- Django
- Node.js
- Kafka
- Redis
- Temporal
- Celery
- RabbitMQ
- AWS
- GCP
- Docker
- Kubernetes
- LangChain
- LangGraph
- Qdrant
- OpenAI
- Claude
- PostgreSQL
- MySQL
- MongoDB
About
I’m a backend engineer with 5+ years of building production systems, and I’ve done it at three very different scales: a large IT services company, a fast-growing commerce startup, and an early-stage AI startup.
At LTIMindtree I worked on enterprise document platforms, where correctness and process mattered most. At Fynd I built order systems for JioMart handling 100k+ orders a day, where the job was staying up under peak load. At EDRA Labs I build AI products from nothing, where speed and judgment matter more than process. Each place taught me something the others couldn’t.
What carries across all three: I like owning a problem end to end, from figuring out what to build, through the architecture and the deploy, to fixing what production tells me. These days that mostly means the backend work behind reliable AI products: orchestration, durable workflows, and async pipelines.
Contact
Open to senior / staff backend roles. Have a role, a problem, or just want to compare notes on systems and AI tooling? My inbox is open.