Headline

Platform Architect | National-Scale Financial Infrastructure | Open Source & DPG | Finternet

About

I am a Platform Architect specializing in designing and scaling national-level financial infrastructure and digital public goods. With a track record spanning 0-to-1 startups and government-scale systems, I excel at building pluggable, modular architectures that prioritize long-term maintainability over quick fixes. My approach is end-to-end; I believe the true quality of an architecture is only verified through production load. Leveraging deep experience in systems design, AI integration, and high-throughput data processing, I help organizations navigate complex scaling challenges and modernize critical technological foundations.

Location

Bengaluru, Karnataka

AI stack

claude code

codex

langchain

langgraph

notebooklm

obsidian

Work experience

Platform Architect at Sanketika, May 2023 - Present · 3 yrs 3 mos, Bengaluru, Karnataka, India, A niche architecture consultancy of 30 people, driving the architecture for national-scale Digital Public Infrastructure and enterprise platforms. Each engagement: formal architecture spec → guide client team → contribute to critical components. - Finternet: Whitepaper to alpha in 6 months. Designed UNITS architecture, built the token engine in Rust with pluggable token programs, orchestration API in Go, full infra on Terraform/ArgoCD. - National Payments Corporation of India: Designed and Built computation pipeline, query engine, and the org-wide modular Data Processing Framework processing about 500M+ transactions per day - Obsrv: Built connector framework (Scala/Python SDKs), multi-cloud infra across AWS/GCP/Azure. Production at Income Tax Dept of India and Infosys. Digital Public Good. - Sunbird: Lead Architect for India's flagship open-source DPI platform, driving all features and global adoption. · Life Insurance Corporation of India: Architected 7-stage real-time CDC pipeline (MySQL→Kafka→MongoDB) for India's largest life insurer (300M+ policies).

Software Architect at Tripaneer, May 2021 - Apr 2023 · 2 yrs, Amsterdam, North Holland, Netherlands, Led architecture and engineering for a distributed European team during hyper-growth. - Architected multi-currency booking engine with payment workflows and accounting integration - Built ML platform: BERT vector search, booking intent model (+10% conversions), page-migration classifier (pruned millions of stale pages), ETL rewrite (60% faster) - Established data engineering and MLOps practices; led AI R&D (dynamic pricing, image captioning, attribution)

Architect & Open Source Contributor at ML Square, Dec 2018 - Dec 2021 · 3 yrs 1 mo, Bengaluru Area, India, - Developed a python package mlsquare to simplify the conversion of standard machine learning models to deep learning models for machine learning developers. - Implemented a pluggable architecture that allows for seamless integration of any machine learning or deep learning library into the package. - Acted as a supervisor for design decisions, upcoming features, and participated in code reviews to ensure high quality and maintainability of the package

Head of Engineering (Founding) at NPBridge, Mar 2018 - May 2021 · 3 yrs 3 mos, Bengaluru Area, India, Joined as Head of Engineering and First Engineer. Built the full stack from scratch: architecture, infrastructure, and engineering practices. Cloud-native on AWS with Kubernetes and Terraform. - Pluggable web crawler: 40K pages/day; source definitions (URL patterns, crawl rules, selectors) were declarative configs, so adding a new domain meant writing an entry to the database, not a new adapter - SaaS vector search engine with NLP pipelines and domain classifiers at 98% accuracy using Transformers and BERT (2018, pre-mainstream)

Member Of Technical Staff at IQLECT, Dec 2016 - Feb 2018 · 1 yr 3 mos, Bengaluru Area, India, - Integrated real-time predictions into the stream processing engine of BangDB (an in-house database) and implemented an LRU mechanism to load models in memory, resulting in a reduction of latency by 80%. - Developed an R Package for BangDB, enabling exploratory analysis using the R ecosystem on data streams. - Created products for Click Stream, eCommerce, Payment Gateways, Mutual Funds, and eLearning, and implemented use-cases where real-time actions needed to be taken, and designed flows for data slicing and dicing. - Created model definitions for Fraud Detection in payments and Engagement Prediction for eLearning apps, which can be directly consumed by customers and fine-tuned on their own data. <inFlight>

Research Engineer at Indian Institute of Technology, Kanpur, Aug 2014 - Dec 2016 · 2 y 4 mo, Kanpur

Education

Master of Computer Science, Arizona State University

Bachelor of Technology (B.Tech.), Computer Science, Jawaharlal Nehru Technological University, Kakinada

Public links

github: https://github.com/ravismula