Headline

AI Software Engineer @ BCG | Full-Stack (Django, React, TypeScript) | GenAI & RAG Pipelines | Open Source (PyPI: docpipe-sdk, vm-tool)

About

AI software engineer with 3+ years of experience building and operating production AI products across Python backends, React/TypeScript interfaces, RAG and agent workflows, and cloud deployment. At BCG X, I work on enterprise GenAI systems including retrieval over 10,000+ slides, agent workflows, evaluation pipelines, reusable AI application scaffolds, and deployments across AWS, Azure, and GCP. My work covers product engineering, CI/CD, monitoring, guardrails, and production reliability. Previously, I was a founding engineer at Auctopus Technologies, where I helped build NetPeace, a network automation and observability SaaS platform managing 1,000+ distributed devices across 15+ containerized microservices. I also build open-source developer tools: • docpipe-sdk — a composable document-to-RAG Python SDK with multiple retrieval strategies • vm-tool — deployment automation for VMs, Docker, Kubernetes, CI/CD, and monitoring using Python and Ansible Core stack: Python, Django, FastAPI, React, TypeScript, LangChain, LangGraph, PostgreSQL, Redis, Docker, Kubernetes, AWS, Azure, and GCP. Based in Bengaluru and open to suitable roles in India, remote teams, and European opportunities with relocation or visa support.

Location

Bengaluru

AI stack

claude code

codex

hugging face

langfuse

langgraph

langsmith

Work experience

AI Software Engineer at BCG X, Jun 2025 - Present · 1 yr 4 mos, Bengaluru, • Engineered MSP B2B Pricing as part of BCG's Helix AI team: ingested 10,000+ slides as a knowledge base and built a RAG-powered chat interface with REST APIs for on-the-fly PowerPoint generation using LLM-driven content synthesis. • Deployed MSP B2B Pricing on Azure Container Apps with Blob Storage, Azure PostgreSQL (pgvector), and Document Intelligence across staging and production environments. • Built SurveySuite, a browser-based platform making Claude Code accessible to non-technical users with isolated sandboxes, guided chat assistants, and integrated terminals for autonomous task execution. • Deployed SurveySuite on AWS EKS via Octopus Deploy and Helm, using ElastiCache (Redis) for Celery broker/caching, S3 for workspace backups, and Bedrock AgentCore for agent infrastructure. • Developed the DFX Boilerplate Template, a modular full-stack AI scaffold allowing developers to configure databases, ingestion pipelines, vector stores, and chaining mechanisms per service, adopted across 10+ BCG projects, reducing setup time by 1 week. • Optimized prompt engineering for Risk Radar, an AI risk analysis tool for a Dubai-based client deployed on GCP (Compute Engine, Cloud SQL, Vertex AI), enabling GPT-5 migration and reducing AI costs by ~70%. • Owned end-to-end deployment across all projects, building CI/CD pipelines, configuring staging and production environments, and managing monitoring across Azure, AWS, and GCP.

Founding Engineer at Auctopus Technologies Pvt Ltd, Mar 2023 - May 2025 · 2 yrs 3 mos, Delhi, India, • Led architecture and end-to-end development of NetPeace, a network automation and observability SaaS platform managing 1,000+ distributed devices across 15+ containerized microservices. • Built scalable REST APIs and real-time WebSocket systems using Django REST Framework and Django Channels, applying efficient data structures and algorithms for live bi-directional communication across Agile sprint cycles. • Engineered React.js frontend with drag-and-drop workflow builder, real-time dashboards, and dynamic form generation, automating 80% of operational processes with 90% decision accuracy. • Automated 100+ Camunda BPMN workflows integrated with ServiceNow, reducing manual incident handling by 90%. • Developed a RAG-powered conversational AI interface using LangChain and vector databases for natural language querying of device configurations, improving troubleshooting efficiency by 35%. • Implemented observability using Prometheus, Grafana, and Alertmanager, improving issue detection by 25% and reducing monitoring effort by 40%. • Owned Stripe payment integration for subscriptions and billing, reducing transaction failures by 25%; optimized Celery and Redis caching, deployed via Docker, Ansible, and GitHub Actions CI/CD, improving platform stability by 20%.

Education

B.Tech, Computer Science, Maulana Abul Kalam Azad University of Technology, West Bengal formerly WBUT

Public links

github: https://github.com/thesunnysinha

sunnysinha.online: https://www.sunnysinha.online/