Harshit Dixit profile photo 1 on Tal

Harshit Dixit

MTS @ Oracle | Vision AI | Model Serving & Inference APIs | LLM & Search Systems | Distributed Systems | Ex-Chargebee | Ex-ZEE5

Bengaluru

4+ years of work experience

Oracle
Aug 20251 yr 1 moPresent

Bengaluru, Karnataka, India

Chargebee
Mar 20241 yr 6 mosAug 2025

Bengaluru, Karnataka, India

• Built AI-powered natural language filter search using LLM + RAG, converting user queries into structured database filters across subscriptions, customers and plans. • Developed schema-aware validation layer ensuring AI-generated filters map correctly to backend database fields before query execution. • Built AI observability pipeline logging prompts, responses and metadata to monitor and debug LLM behaviour in production. • Implemented large-scale OpenSearch data cleanup workflows using Delete-By-Query APIs with concurrency control and rate limiting. • Optimized production systems through SQL query tuning, indexing and monitoring using Datadog and Splunk.
ZEE5
Nov 20235 mosMar 2024

Bengaluru, Karnataka, India

• Implemented Vertx eventBus to re-architecture monolithic content APIs into low-latency, reactive microservices, handling 4k transactions per second (TPS). • Developed and optimized high-traffic APIs, reducing error incidents by 15% through robust error handling and logging mechanisms. • Introduced Redis caching system, decreasing latency by 25X (from 250ms to 12ms) for high-traffic content APIs using Java Vert.x. • Utilized Kafka for real-time messaging, orchestrating cache purging upon new asset publications on Zee5 platform. • Successfully deployed APIs and services on AWS, migrating select services to GCP for optimal scalability and resource utilization.
ZEE5
Aug 20221 yr 2 mosSep 2023

• Developed a Spring Boot-powered system for dynamic asset sorting based on priority rules, enhancing user experience and engagement. • Designed and implemented a Spring Boot application for parsing Excel sheets and extracting asset details, integrating with MongoDB and external APIs for real-time recommendations. • Utilized MongoDB integration for retrieving detailed asset information and Spring Boot's RESTful capabilities for communication with external recommendation APIs. • Increased content relevancy by 30% through dynamic rule-based sorting algorithms, resulting in a noticeable reduction in the page leaving rate by users.
ZEE5
Apr 20225 mosAug 2022

Harshit’s AI Stack

anthropic apis
anthropic apis
cursor
cursor
kimi
kimi
langchain
langchain
langfuse
langfuse
langgraph
langgraph

Education

Bms institute of technology and management

Bachelor of Technology - BTech, Computer Science Engineering

Bms institute of technology and management

20184 yrs2022

X and XII, PCM

St. Anthony's senior secondary school, Fatehgarh

20172017

Join Harshit and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Harshit Dixit on Tal

Headline

MTS @ Oracle | Vision AI | Model Serving & Inference APIs | LLM & Search Systems | Distributed Systems | Ex-Chargebee | Ex-ZEE5

About

🚀 MTS @ Oracle | AI Vision Systems | AI Platforms | MLE , Backend & Cloud Infrastructure 📈 ⏳ YOE: 4 | 🧠 Lifelong Learner 🌟 About Me 🌟 Currently part of the Oracle Cloud Infrastructure (OCI) Vision AI team, where I work on production systems that serve computer vision models through scalable APIs. Our platform enables developers to deploy and run AI models via synchronous and asynchronous workflows, manage projects, and run custom models at scale. My work involves integrating trained models from research teams into reliable production systems — building AI inference pipelines, backend services, and cloud workflows using technologies like Docker, Kubernetes, and Terraform. Previously at Chargebee, I was part of a team dedicated to enhancing platform infrastructure, optimizing performance, and ensuring scalability and reliability for SaaS solutions. Before that at ZEE5, I focused on developing innovative solutions for OTT platforms, which deepened my expertise in software development and problem-solving. 🔧 My Expertise 🔧 Chargebee: Platform Infrastructure: Assisting in the development and maintenance of the Chargebee platform, focusing on scalability and reliability. Integration & Performance: Learning to handle integrations with third-party systems and working on performance optimization. ZEE5: RCS Service Development: Contributed to dynamic sorting and personalized content recommendations using Spring Boot and Vert.x, enhancing relevancy by 30%. Content API Transformation: Improved API performance with Java Vert.x, achieving 25X latency reduction, and optimized deployments with Redis and Kafka on AWS and GCP. 🚀 Let’s Connect! 🚀 I’m keen to connect with fellow tech enthusiasts, industry professionals, and innovators. Whether it’s about technology trends, potential collaborations, or sharing experiences, let’s connect and explore new opportunities together! Email: harshit.dixit2022@gmail.com

Location

Bengaluru

AI stack

anthropic apis

cursor

kimi

langchain

langfuse

langgraph

Work experience

Member of Technical Staff at Oracle, Aug 2025 - Present · 1 yr 1 mo, Bengaluru, Karnataka, India

Software Engineer at Chargebee, Mar 2024 - Aug 2025 · 1 yr 6 mos, Bengaluru, Karnataka, India, • Built AI-powered natural language filter search using LLM + RAG, converting user queries into structured database filters across subscriptions, customers and plans. • Developed schema-aware validation layer ensuring AI-generated filters map correctly to backend database fields before query execution. • Built AI observability pipeline logging prompts, responses and metadata to monitor and debug LLM behaviour in production. • Implemented large-scale OpenSearch data cleanup workflows using Delete-By-Query APIs with concurrency control and rate limiting. • Optimized production systems through SQL query tuning, indexing and monitoring using Datadog and Splunk.

Software Engineer 1 at ZEE5, Nov 2023 - Mar 2024 · 5 mos, Bengaluru, Karnataka, India, • Implemented Vertx eventBus to re-architecture monolithic content APIs into low-latency, reactive microservices, handling 4k transactions per second (TPS). • Developed and optimized high-traffic APIs, reducing error incidents by 15% through robust error handling and logging mechanisms. • Introduced Redis caching system, decreasing latency by 25X (from 250ms to 12ms) for high-traffic content APIs using Java Vert.x. • Utilized Kafka for real-time messaging, orchestrating cache purging upon new asset publications on Zee5 platform. • Successfully deployed APIs and services on AWS, migrating select services to GCP for optimal scalability and resource utilization.

Associate Software Development Engineer at ZEE5, Aug 2022 - Sep 2023 · 1 yr 2 mos, • Developed a Spring Boot-powered system for dynamic asset sorting based on priority rules, enhancing user experience and engagement. • Designed and implemented a Spring Boot application for parsing Excel sheets and extracting asset details, integrating with MongoDB and external APIs for real-time recommendations. • Utilized MongoDB integration for retrieving detailed asset information and Spring Boot's RESTful capabilities for communication with external recommendation APIs. • Increased content relevancy by 30% through dynamic rule-based sorting algorithms, resulting in a noticeable reduction in the page leaving rate by users.

SDE Intern at ZEE5, Apr 2022 - Aug 2022 · 5 mos

Education

Bachelor of Technology - BTech, Computer Science Engineering, Bms institute of technology and management

X and XII, PCM, St. Anthony's senior secondary school, Fatehgarh