Vivek Keshava

Senior Software EngineerDistributed SystemsAI Tooling

I build high-throughput backend systems. Recently built and scaled a used-car marketplace, and I'm building AI developer tooling with LLMs and the Model Context Protocol.

GitHub LinkedIn Email

About

I'm a senior software engineer with 5+ years designing high-throughput backend systems, with a focus on system design, reliability, and performance at scale. Most of my work lives in event-driven microservices and distributed platforms built with reactive Java, Apache Kafka, AWS, and Kubernetes.

At Credit Acceptance I led the architecture of a used-car marketplace that grew from 2,000 to over a million daily users, owning everything from credit-application workflows to async notification infrastructure handling 1M+ daily deliveries. Lately I've been building AI-powered developer tooling using LLMs and the Model Context Protocol.

Work Experience

Senior Software Engineer

Credit Acceptance Corporation

December 2025 - CurrentPhoenix, USA (Remote)

• Led end-to-end architecture of a high-traffic used-car marketplace, scaling from 2,000 to 1,000,000 daily users across credit-application workflows.
• Designed a Redis cache-aside strategy with TTL-based invalidation across high-traffic read paths, cutting P95 API latency by ~35%; built concurrent, non-blocking services with Spring WebFlux across 12+ microservices.
• Designed and delivered asynchronous notification services on AWS handling 1M+ daily deliveries with ~40% latency reduction.

Software Engineer II

Credit Acceptance Corporation

January 2024 - December 2025Phoenix, USA (Remote)

• Led a zero-downtime migration from gRPC + Micronaut to REST + Spring Boot WebFlux across 8 interdependent services under live traffic, improving throughput 30% and cutting P50 latency 20%.
• Architected cloud-native microservices on Kubernetes & Helm (AWS EKS, S3, Lambda), achieving 99.9% uptime and reducing deployment cycles 25% through CI/CD automation.
• Built an end-to-end payment platform with Apollo GraphQL federation (NestJS) and a custom OAuth 2.0 token system for guest payments, supporting thousands of daily transactions with zero auth downtime.

Software Engineer II

Micro Focus

Bengaluru, India

• Architected reusable data streaming pipelines with Apache Kafka and Apache Pulsar across 5 products, improving throughput 30% and reducing development cycle time 20%.
• Developed high-performance REST APIs for the NNMi backend processing data from 100+ network devices, powering monitoring dashboards for 500+ enterprise customers.
• Architected and migrated the on-premises NOM product to AWS with a 7-member team in 3 months, resulting in a 40% reduction in operational costs.

Software Engineer I

Micro Focus

• Designed reusable data streaming components and distributed system frameworks used across 5 products, improving data processing speed 30%.
• Developed multiple REST APIs for the NNMi backend framework, enabling analytics on network data from 100+ devices.
• Built web applications and services used by 500+ customers worldwide, increasing customer satisfaction 40%.

Technical Skills

Languages

Java

Python

C++

JavaScript

TypeScript

Databases

Vertica

MySQL

PostgreSQL

Neo4j

Oracle DB

Cloud & DevOps

AWS

Kubernetes

Docker

CI/CD

EKS

Lambda

Web & Frameworks

React.js

Spring Boot

GraphQL

REST API

Node.js

NestJS

Education

Master of Science in Computer Science

Arizona State University - Tempe, Arizona

December 2023

GPA: 4.0/4.0

Courses: Foundation of Algorithms, Database Management and System Implementation, Statistical Machine Learning, Mobile Computing, Data Mining, Data Processing at Scale, Data Visualization

Bachelor of Engineering, Electronics and Communication

Sri Jayachamarajendra College of Engineering - Karnataka, India

May 2018

GPA: 9.08/10

Courses: Data Structures and Algorithms, Computer Concepts and C Programming, Networking, Embedded Systems, Operating Systems

Featured Projects

GitHub Engineering Intelligence MCP Server

TypeScript

MCP

LLMs

Production-ready MCP server built with the TypeScript MCP SDK, Zod validation, and the GitHub REST API, exposing repository intelligence as structured tools for AI agents and LLMs. Implements repo health scoring, PR risk classification, and workflow bottleneck detection — deployed as a live integration in Claude's ecosystem.

Technologies: MCP SDK, GitHub REST API, Zod, TypeScript, Python

View on GitHub

LLM-based Document Parser & Authenticator

Python

LLMs

React

LLM-powered parser that extracts structured data from unstructured text, validates authenticity, and flags anomalies with over 90% accuracy. Integrated open-source models (Llama, DeepSeek) for extraction and accuracy scoring, cutting manual verification by 60%.

Technologies: Llama, DeepSeek, Python, TypeScript, React

Stance Detection on Twitter Data

Python

Machine Learning

Engineered and trained ML models for stance detection on Twitter data using SVM, RNN, and LSTM to classify a reply's position relative to the source tweet. Achieved a 10% accuracy improvement through hyperparameter tuning.

Technologies: Scikit-Learn, NumPy, Pandas, SVM, RNN, LSTM

Bi-directional Stock Prediction

Python

Deep Learning

Binary classifiers that predict stock trends from sentiment analysis of finance news and time-series market data. Benchmarked traditional ML (SVM, random forest, logistic regression) against deep learning (LSTM, XGBoost) with feature selection and grid-search tuning.

Technologies: Keras, XGBoost, LSTM, Scikit-Learn

Publications

Research Publication

Robotic Mapping Using Autonomous Vehicle

Keshava, Vivek, et al. "Robotic Mapping Using Autonomous Vehicle." SN Computer Science, vol. 1, no. 3, May 2020

View Publication

2025 Reading List

A few books I enjoyed this year

The Library of Borrowed Hearts — Lucy Gilmore
How to Stop Time — Matt Haig
Atomic Habits — James Clear
The Kite Runner — Khaled Hosseini

Let's Connect

I'm always interested in discussing new opportunities, innovative projects, and collaborations.

vivek.keshava1@gmail.com