@shikhar13012001

ABOUTWORK

WELCOME TO THE  

WORKSPACE OF 

SHIKHAR 

DISCOVER. ME.

About




I'm a Software Engineer at American Express, based in Bengaluru, India. I build AI-driven products, agentic developer tooling, and backend systems.


At American Express I work on Automation & Tools — shipping an LLM-powered VS Code extension for autonomous test generation, a terminal-native package manager for our internal Claude skills ecosystem, and an ML-based record-linkage system for payment adjudication. Before that, I built a RAG platform at Nasdaq and scaled a 200k-user product at Futurepedia.


Outside of work, I build my own AI agents and voice systems end to end — from a multilingual clinic voice receptionist to a multi-agent video-generation pipeline — because I like owning a product from architecture through to the last mile of the UI.

FIND MY RESUME HERE

shikhar13012001

Work Experience

Where I've built software professionally, most recent first.

(01)

American Express

Software Engineer — Automation & Tools

Bengaluru, India · January 2025 – Present

Built a VS Code extension that autonomously generates JUnit tests for Java codebases using LLMs, AST parsing, and build-validate-fix loops — integrating JaCoCo coverage analysis and automated retry logic, cutting test-creation time by ~20%.

Built Skills-PM, a terminal-native package manager for Amex's internal Claude skills ecosystem (TypeScript, Ink, Octokit) with multi-registry federation and a lock-file system, enabling discovery and install of 15,000+ skills without redundant cloning across 12+ teams.

Designed a record-linkage system for France-market bill-payment adjudication — an 8-stage pipeline pairing Fellegi-Sunter probabilistic scoring with a calibrated LightGBM classifier over a 25-feature name-comparison vector, served via a Flask REST API.

Built automated PR review agents on OpenCode and a Confluence-to-video documentation pipeline (Remotion + Claude); handled Dependabot, EarlyBird secrets-scanning, and OpenTelemetry instrumentation work across services.


(02)

Nasdaq

Software Engineering Intern

Remote · June 2024 – July 2024

Developed a RAG-based conversational agent backend for large, sensitive documents using Spring Boot and Azure OpenAI, improving processing accuracy by ~30%.

Designed and implemented a parallel processing framework for batch embedding creation, cutting document indexing time by ~80%.

Built a prototype iOS app with an intuitive UI, integrating the backend via Alamofire.


(03)

Futurepedia

Software Engineering Intern — Full Stack

Remote · October 2022 – June 2024

Scaled a Next.js + Sanity CMS platform to 200,000+ users on Vercel.

Led integration of Typesense vector search, improving search precision by ~60%.

Added serverless caching with Upstash Redis and custom lazy-loaded YouTube embeds — cutting operational costs by ~19%, lifting the Lighthouse score by ~20 points, and improving conversions by ~15%.



Tools / Libraries
There are many tools out there for backend systems, AI pipelines and developer tooling. These are the ones I reach for most. In my opinion tools are not permanent and I'll always be trying out new stuff :)
Java

01

Java
Primary language for production backend services and JVM tooling at American Express
Spring Boot

02

Spring Boot
Framework for production backend services and internal developer tools
Python

03

Python
Language of choice for ML pipelines, model-scoring APIs, and automation tooling
Next js

04

Next js
React framework for full-stack product frontends
TypeScript

05

TypeScript
Typed JavaScript I default to for frontends, CLIs and Node services
LangChain

06

LangChain
Agentic pipelines and retrieval-augmented systems on GPT-4o, Claude and Azure OpenAI
Docker

07

Docker
Containerizing services for consistent local dev and deployment
Kubernetes

08

Kubernetes
Orchestrating containerized workloads in production
PostgreSQL

09

PostgreSQL
Primary relational store for application and clinical data via Supabase and Prisma
Twilio

10

Twilio
SIP telephony and messaging infrastructure behind my voice AI and WhatsApp integrations
GitHub Actions

11

GitHub Actions
CI/CD pipelines for build-validate-test-deploy workflows
Vercel

12

Vercel
Deployment platform for Next.js apps and preview environments
Still learning new stuff, I’ll be adding more tools as I learn new :)

Projects

A mix of production systems, AI agents, and independent builds. Some of my work at American Express is internal and isn't publicly linked, so it isn't listed here.

(01)

ScheduRx

ScheduRx

(02)

PostMortemEng

PostMortemEng

(03)

Rampify

Rampify

(04)

Futurepedia

Futurepedia

Published Packages

These are the packages that I have published on npm. They are all open source and free to use. I have built these packages to make my life easier and I hope they will help you too.

CV-Letter

CV-Letter

cv-letter is a Command Line Interface (CLI) tool that combines a resume, a job description, and a LaTeX template to create a personalized AI-generated cover letter. It utilizes OpenAI's GPT-3.5 model to tailor the content to specific job applications, highlighting crucial details like email contact, company names, and relevant skills.


cv-letter

Minal-CLI

Minal-CLI

MonkeyType Terminal is a powerful, open-source typing practice tool designed specifically for the command-line interface. It offers an immersive and interactive way to improve your typing speed and accuracy right from your terminal.Experience the thrill of improving your typing skills with MonkeyType Terminal.


monk-minal

flower

Let's get in touch

I'm always open to interesting conversations and opportunities. Looking for a SOFTWARE ENGINEER with AI and backend experience? Just get in touch.