Aakarsh Pattanaik

Designed and owned an end-to-end order automation service using FastAPI, building a modular two-stage workflow (site → product) with persistent state management that cut manual operations by 60–70%. Led the architecture and low-level design of a distributed document processing platform, orchestrating 10+ independent agents with well-defined contracts and shared data models to drive a 3–4x increase in investigation throughput. Built and deployed a scalable AWS-based data ingestion and scraping pipeline for web and mobile sources, optimizing execution flow and resource utilization to shrink end-to-end processing time from 40 minutes to 2 minutes for Android app–based scraping.

Owned and operated AWS infrastructure (EC2, security groups, VPNs) across 5+ production and development environments supporting 30+ users, reducing environment setup time by 50% and improving onboarding reliability and system stability. Designed and built Python-based data processing services to transform and aggregate 100k+ records into structured PDF outputs, cutting manual data preparation effort by 60% through automated pipelines.

Built and maintained NLP-based text processing pipelines for semantic extraction across 10k+ documents, improving downstream model precision by 18% and reducing manual validation effort by 40%. Built a document-driven recommendation system using RAG over 1k+ loan profiles, enabling personalized borrowing plans and reducing recommendation turnaround time by 30%.

Bhopal, Madhya Pradesh