Data science · engineering · applied AI
I’m Sanxchep, a data engineer and scientist turning research, machine learning, and difficult infrastructure into systems people can actually depend on.
01 / Selected systems
/01
A Django OAuth 2.0 and IAM service with access and refresh token flows, RFC 7662 introspection, client-to-resource authorization, Alexa sign-in, health checks, and production deployment tooling.
02 / Project archive
Most of this work shipped as connected product services, not isolated demos. The archive groups those service families by the problems they solved.
A multi-application Django authentication service with token issue and refresh flows, RFC 7662 introspection, resource authorization, Alexa login, and IAM APIs.
Face enrollment and recognition connected to attendance workflows, AWS services, REST endpoints, schedulers, logs, test collections, and production packaging.
A service family spanning transformer inference, question generation, data pipelines, API orchestration, sample tooling, and batch-parallel inference across eight GPUs.
Chatbot proofs of concept and service backends built across GPT-3, GPT-3.5, GPT-4, Pinecone retrieval, streamed responses, and reusable assistant APIs.
Alexa skills, Lambda handlers, Lex v2 endpoints, intent routing, authentication, response helpers, and backend services for the Genie assistant.
Reusable Django foundations across multiple product teams and releases.
Focused text-search and nudge-delivery APIs built for a larger education platform.
Focused builds for Amazon product categorization, a blockchain proof of concept, and a Hyperledger Fabric client—kept as supporting work rather than headline projects.
Deployment scripts, Django migrations, ECS tests, cron services, health checks, log handling, and Linux startup tooling used to keep services moving between environments.
03 / Experience
2023 — NOW
August-Wilhelm Scheer Institut · Saarbrücken
Research data infrastructure, provenance, automation, and machine-learning systems for electrochemical manufacturing and testing.
2022 — 2023
Nimoy IT Solutions · Gurugram
Production NLP, ML-backed APIs, OAuth foundations, cloud services, and retrieval-augmented AI prototypes across multiple domains.
2021 — 2022
Nimoy IT Solutions · Gurugram
Dockerized ML services, Django REST APIs, Linux production systems, and an attendance platform powered by face recognition.
04 / Research
05 / External proof
End-to-end backend ownership, scalable delivery, and a consistent eye for inefficiency.
Could turn ideas into proofs of concept in days while handling multiple scalable products.
Fresh approaches, reliable delivery, and unusual commitment when building scalable ML solutions.
07 / Contact