đź‘‹

Hi, I'm Manoj Mishra

Principal engineer who still loves shipping—Rails, React, and AI that actually ships.

I've spent years on high-traffic monoliths and fintech rails (literally)—now I care most about boring reliability + sharp AI tooling. If you want someone who reads diffs and docs with equal enthusiasm, we'll get along.

What I'm exploring

I believe the most interesting engineering today happens where mature Rails systems meet LLMs—not as a rewrite, but as careful augmentation: retrieval, guardrails, and tools that fit how teams already ship. Right now I'm deep on RAG patterns, vision APIs for structured extraction, and developer-facing extensions that make AI feel boringly reliable.

About

I'm a principal engineer based in Noida, Uttar Pradesh. The résumé bullets live in Experience—here's the short version: I like hard problems on mature stacks, clear ownership, and teams that still care about tests when nobody is watching.

Lately that means RAG and LLM integrations alongside Rails and React, cutting AWS spend without drama, and mentoring so code review stays kind and useful. I ship side projects and OSS when an idea won't leave me alone—see Projects for the ones I'm proudest of.

Outside tickets and PRs: I read release notes, tinker with extensions, and write notes when a pattern feels worth writing down—even if it's just a few paragraphs.

Skills

Languages

RubyNode.jsJavaScriptPythonGoJavaC++C

Frameworks & Frontend

RailsNext.jsNestJSExpressJSDjangoPhoenixReactVue.jsAngularjQuery

Databases & Storage

MySQLPostgreSQLMongoDBRedisElasticsearchVector Databases (Chroma)

DevOps & Infrastructure

DockerAWS (EC2, ECS, S3)AKS (Azure Kubernetes Service)GitHub ActionsCI/CDNginx

Queues & Workers

SidekiqResqueDelayed JobRabbitMQBullMQ

AI/ML

Hugging Face TransformersLLM Fine-TuningRAG PipelinesEmbeddings-based SearchLangChainChromaGradioOpenRouter

Experience

Formal employment—roles, scope, and outcomes. For independent builds and OSS, jump to Projects.

Principal Engineer

Bold Technologies (FlexJobs)

Sep 2023 – Present

  • Led the modernization and optimization of a high-traffic Rails 7 monolith, achieving 99.9% uptime for 50,000+ job seekers
  • Reduced AWS infrastructure costs by 35% through strategic Docker containerization and resource tuning
  • Mentored a team of 5, strengthened TDD culture (RSpec/Jest), reducing bugs by 30% and increasing release velocity by 25%
  • Achieved 65% faster API responses through MySQL tuning and Cloudflare caching
  • Built premium job-filtering UX using Vue.js, boosting application conversions by 18%
  • Designed and implemented RAG-based internal tools using OpenAI embeddings and Chroma vector DB, enabling semantic search and automated content retrieval

Tech: Rails, Next.js, Vue.js, MySQL, Docker, AWS S3/ECS

Full Stack Engineer

SpacedVentures

Apr 2020 – Apr 2023

  • Architected a high-security React and Rails fintech platform that processed $500M+ in investment capital for 10,000+ investors with zero security or operational incidents over three years
  • Integrated FundAmerica & NorthCapital APIs; cut transaction processing time by 45%
  • Achieved 85% code coverage via RSpec + Jest, reducing production bugs by 30%

Tech: Rails, React, PostgreSQL, Redux, Docker, AWS

Full Stack Engineer

Kognitive Corporation (Crowdmatrix)

Apr 2017 – Mar 2022

  • Built investor platform enabling $100M+ in diversified allocations
  • Implemented TDD pipelines (RSpec/Jest) reaching 90% coverage, reducing defect leakage by 40%

Tech: Rails, React, MySQL, Docker, AWS

Team Lead

Josh Software Pvt Ltd (Brandscope)

Jul 2013 – Apr 2016

  • Led team of 8 to deliver a fashion e-commerce platform increasing checkout conversions by 40%; drove $10M+ annual sales
  • Introduced engineering standards improving code quality by 50%
  • Created multi-tenant SaaS architecture supporting 10 clients, cutting deployment complexity by 60%

Tech: Rails, jQuery, MongoDB, Nginx, Passenger, AWS

Software Engineer

Josh Software Pvt Ltd (Sprycer)

Apr 2012 – Jun 2013

  • Built large-scale price comparison engines with crawling, scraping, and analytics layers
  • Adopted TDD using RSpec for reliability and correctness

Tech: Rails, MongoDB, Elasticsearch, Redis, Angular.js, AWS

Side projects & open source

Things I build on my own time—tools, experiments, and small products. For day-job impact and timelines, see Experience.

Featured work below; the rest lives on GitHub with the same care for READMEs and reproducible setup.

đź’ˇ

InsureLLM RAG Assistant

RAG chatbot with LangChain, Chroma, and Gradio—semantic retrieval and Q&A over domain documents. A playground for grounded answers and vector search patterns I reuse in production tooling.

LangChainChromaGradioPythonRAG
View on GitHub
đź§ľ

Receipt Extractor

Full-stack receipt parsing: NestJS API, BullMQ workers, React/Vite UI, and OpenRouter vision models. Upload an image, get structured vendor, line items, and totals—real-world LLM extraction with queues and storage.

NestJSReactPrismaBullMQOpenRouter
View on GitHub
🤖

Bitbucket PR Review Extension

Chrome extension that brings GPT-powered review to Bitbucket pull requests—fast feedback on diffs without leaving the host. Built for teams that live in Bitbucket, not GitHub.

Chrome ExtensionJavaScriptBitbucket APIGPT
View on GitHub

More repos

🚀

Grok Everywhere

Grok in the browser—Chrome extension for Grok API access across the web.

Chrome ExtensionJavaScript
GitHub →
🔍

Sourcemap Decoder

Decode sourcemaps and map minified stack locations back to original source.

Node.jsCLInpm
GitHub →
▶️

Launchpad

VS Code extension: Procfile-style processes, one terminal, Foreman-style prefixed logs.

VS CodeTypeScriptProcfile
GitHub →

Notes

Short write-ups—things I've learned building with AI and shipping serious money paths. Full long-form posts may follow; for now, here's how I think in public.

AI / infra

How I think about RAG for internal docs

Start with the question your coworkers actually ask, not the embedding model leaderboard. Chunking, metadata, and a boring evaluation loop beat clever prompts. At work we paired OpenAI embeddings with Chroma and tight feedback from editors—semantic search only matters if humans trust the first page of results.

Fintech

Lessons from $500M moving through a Rails stack

Money teaches you that “it works on my machine” is never enough. Idempotency keys, audit trails, and treating third-party APIs as flaky by default matter more than framework debates. The win was boring: exhaustive tests, clear ownership on incidents, and never shipping a payment path you couldn’t explain to compliance.

Opinion

Rails + LLMs: augment, don’t rip out

Monoliths that already handle auth, billing, and domain rules are an asset. I’d rather wire an LLM behind a small service boundary—with Zod-shaped outputs and human review—than chase a greenfield rewrite. The intersection I care about is product velocity with safety rails, not replacing Ruby with buzzwords.

Get In Touch

I'm always open to discussing new projects, creative ideas, or opportunities to be part of your visions.

This form is wired to Formspree—submissions go straight to my inbox. Prefer email? Use the link below.