Yash Salunke profile photo 1 on Tal

Yash Salunke

SWE @Pubmatic | Former SDE Intern @PTC | Aspiring Java Developer | Tech Enthusiast | CS’24

Pune Division

3+ years of work experience

PubMatic
Jul 20242 yrs 1 moPresent

Pune District, Maharashtra, India · Hybrid

Project & Contributions ● Real-Time Bid Rejection Troubleshooting Pipeline Built a real-time analytics and troubleshooting pipeline using Apache Kafka and Spark Streaming, processing 100K+ bid events/sec with ~92% accuracy. Delivered near real-time (2-minute batch) insights via Snowflake and enhanced observability using InfluxDB and Grafana, reducing incident detection time by ~40%. ● Buyer-Side Reporting Pipeline Developed a buyer-side reporting pipeline for winning impressions using batch processing, storing data in HDFS and Snowflake to enable scalable and efficient reporting. ● Fluid Timezone Support Implemented dynamic timezone-aware reporting capabilities, enabling accurate, localized reporting across global timezones and significantly improving user experience. ● Redis Sentinel Cluster on Kubernetes Led a POC to set up a Redis Sentinel-based distributed caching system on Kubernetes and implemented it across data access services, reducing API response time by ~30–40% and improving system scalability and resilience. ● GenAI-Based Ad Inventory Recommendation System Designed and deployed a GenAI-powered recommendation system using K-Means clustering, RAG architecture, and hybrid search (BM25 + vector embeddings on Milvus), improving campaign efficiency by 35% and recommendation relevance by ~89%. ● GenAI Service Optimization Optimized FastAPI-based GenAI services using asynchronous processing, Redis caching, and prompt persistence, reducing latency by ~45% and achieving ~8 seconds average response time.
PubMatic
Jan 20246 mosJul 2024

Pune District, Maharashtra, India · Hybrid

● Observability Tool Contributed to an observability platform for monitoring Apache Spark jobs, capturing key execution metrics such as runtime, executor usage, and cost, enabling performance optimization and cost analysis.
Onshape, a PTC Technology
Jul 20236 mosDec 2023

Kalyani Nagar

GitHub
@Yas18
LeetCode
@Yas_18

Yash’s AI Stack

claude code
claude code
cursor
cursor
hugging face
hugging face
langfuse
langfuse
langgraph
langgraph
openclaw
openclaw

Education

Pimpri Chinchwad Education Trust'S. Pimpri Chinchwad College Of Engineering

Bachelor of Technology - BTech, Computer Engineering

Pimpri Chinchwad Education Trust'S. Pimpri Chinchwad College Of Engineering

20204 yrs2024

Rayat Shikshan Sansthas R.B.Narayanrao Borawake College, Shrirampur, Dist.Ahmednagar 413709

Join Yash and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Yash Salunke on Tal

Headline

SWE @Pubmatic | Former SDE Intern @PTC | Aspiring Java Developer | Tech Enthusiast | CS’24

About

Software Engineer with 2+ years of experience in Java development and Big Data Engineering building scalable, real-time data systems in the AdTech domain. Expertise in Apache Spark, Kafka, and Snowflake for processing high-volume data streams (100K+ events/sec) and delivering near real-time insights. Proven track record in designing high-performance backend systems using Java (Spring Boot), improving system observability, and optimizing distributed architectures for scalability and cost efficiency. Recently focused on GenAI-driven solutions, building recommendation systems using RAG and hybrid search to enhance campaign efficiency and user experience.

Location

Pune Division

AI stack

claude code

cursor

hugging face

langfuse

langgraph

openclaw

Work experience

Software Engineer at PubMatic, Jul 2024 - Present · 2 yrs 1 mo, Pune District, Maharashtra, India · Hybrid, Project & Contributions ● Real-Time Bid Rejection Troubleshooting Pipeline Built a real-time analytics and troubleshooting pipeline using Apache Kafka and Spark Streaming, processing 100K+ bid events/sec with ~92% accuracy. Delivered near real-time (2-minute batch) insights via Snowflake and enhanced observability using InfluxDB and Grafana, reducing incident detection time by ~40%. ● Buyer-Side Reporting Pipeline Developed a buyer-side reporting pipeline for winning impressions using batch processing, storing data in HDFS and Snowflake to enable scalable and efficient reporting. ● Fluid Timezone Support Implemented dynamic timezone-aware reporting capabilities, enabling accurate, localized reporting across global timezones and significantly improving user experience. ● Redis Sentinel Cluster on Kubernetes Led a POC to set up a Redis Sentinel-based distributed caching system on Kubernetes and implemented it across data access services, reducing API response time by ~30–40% and improving system scalability and resilience. ● GenAI-Based Ad Inventory Recommendation System Designed and deployed a GenAI-powered recommendation system using K-Means clustering, RAG architecture, and hybrid search (BM25 + vector embeddings on Milvus), improving campaign efficiency by 35% and recommendation relevance by ~89%. ● GenAI Service Optimization Optimized FastAPI-based GenAI services using asynchronous processing, Redis caching, and prompt persistence, reducing latency by ~45% and achieving ~8 seconds average response time.

Software Engineer Intern at PubMatic, Jan 2024 - Jun 2024 · 6 mos, Pune District, Maharashtra, India · Hybrid, ● Observability Tool Contributed to an observability platform for monitoring Apache Spark jobs, capturing key execution metrics such as runtime, executor usage, and cost, enabling performance optimization and cost analysis.

Java Developer at Onshape, a PTC Technology, Jul 2023 - Dec 2023 · 6 mos, Kalyani Nagar

Education

Bachelor of Technology - BTech, Computer Engineering, Pimpri Chinchwad Education Trust'S. Pimpri Chinchwad College Of Engineering

Rayat Shikshan Sansthas R.B.Narayanrao Borawake College, Shrirampur, Dist.Ahmednagar 413709

Public links

github: https://github.com/Yas18

leetcode: https://leetcode.com/Yas_18