Jay Joshi profile photo 1 on Tal

Jay Joshi

SDE 3 | Data+AI at CommerceIQ | Payments at BukuWarung (YC20 - Fintech) | Search & Recommendation at Blibli

Bengaluru

7+ years of work experience

CommerceIQ
Sep 20233 yrsPresent

Bengaluru

I own and scale CommerceIQ's core data orchestration platform—an Airflow-like system driving 300+ production workflows across 5 engineering teams on Databricks. Originally a basic SQL executor, I led its evolution into a robust, multi-language engine (supporting SQL, PySpark, Scala, and DBT) that is now the company standard for running data pipelines. Lately, I’ve been focusing on fixing bottlenecks in building AI agents and making the platform smarter: - Drastically Cut On-Call Stress: Built an internal RAG support assistant trained on 18 months of incident history and integrated it directly into our IDEs. This slashed mean time to resolution by 88% (from over 2 hours down to 15 minutes), giving the team massive breathing room. - Streamlined Dataset Onboarding: Designed a human-in-the-loop, agentic onboarding tool (via Claude Code) that compressed a painful, multi-day, 5-stage setup process into a single guided session—currently live and onboarding major retail customers. - Saved $200K+/Year in Infrastructure: Built an LLM-assisted SQL-to-Spark migration toolchain alongside workflow optimizations, instantly knocking over $200K/year off our Databricks compute bill. - Boosted Platform Stability & Scale: Anticipated a looming Databricks concurrency bottleneck to ship a multi-workspace routing architecture three weeks ahead of a major breach (tripling capacity headroom). Fixed recurring API throttling with Redis distributed rate limiting, and authored the company's distributed tracing standard. - Team Leadership: Serve as a design-review anchor and mentor for the team. Wrote the platform's support SLA and triage policies, which cut inbound ticket volumes by 50% year-over-year. Promoted to SDE 3 for driving platform architecture, proactive risk mitigation, and proven cost/reliability engineering.
BukuWarung
Apr 20232 mosMay 2023

Bengaluru

Joined the team owning post-payment fulfillment for digital goods (bills, top-ups). Built the core order lifecycle and vendor integration layer, handling real-time delivery, async retries, and automatic vendor fallback routing to ensure reliable distributed system performance. - Emergency Payment Gateway Task Force: Selected for a high-pressure team to build an internal gateway from scratch in 30 days after our provider pulled out. Personally owned the low-level design, crafting an extensible data model for rapid bank onboarding and building disbursement monitoring with automated failovers. - Improved System Reliability & Testing: Refactored synchronous payment flows to async, eliminating timeout-related failures. Built a configurable mock service that replicated unpredictable bank sandbox behaviors, cutting QA cycle times in half. - Standardized Engineering Practices: Designed and drove adoption of a consistent, extensible API response framework used across all backend and frontend teams.
BukuWarung
Sep 20211 yr 8 mosApr 2023

Bengaluru, Karnataka, India

Blibli
Jul 20192 yrs 3 mosSep 2021

Bengaluru

Worked across search infrastructure, distributed systems, and applied ML for one of Indonesia's largest e-commerce platforms. - Led GCP & Kafka Migration: Containerized Java/Spring Boot services using Docker, integrated Vault and Consul, and moved product updates from REST calls to a Kafka-based, event-driven pipeline. Designed an exactly-once processing mechanism (Redis TTL + Postgres row-level locking) to keep data consistent across pods—completing the migration with zero rollbacks and beating the downtime schedule. - Fixed Solr Bottlenecks (10x Throughput): Diagnosed a major performance drag where flash sales and bulk stock refreshes triggered full-document rewrites on massive, 50-field Solr documents. Rewrote the system around Solr atomic partial updates, delivering a 10x improvement in update throughput under peak loads of 12,000 QPS. - Shipped Blibli’s First Live ML Inference Service: Independently built an NLP system to extract product attributes from search queries across a 20M+ catalog. Shipped the model-serving Python API using optimized TensorFlow (quantization/pruning), adding just ~40ms to p99 search latency while cutting clicks-to-"Add to Cart" by 20-25% in major categories.

Jay’s AI Stack

antigravity
antigravity
codex
codex
cursor
cursor

Education

Nirma University

Bachelor's degree, Information Technology

Nirma University

20154 yrs2019

Join Jay and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Jay Joshi on Tal

Headline

SDE 3 | Data+AI at CommerceIQ | Payments at BukuWarung (YC20 - Fintech) | Search & Recommendation at Blibli

About

Experienced backend software engineer with expertise in Java. Worked with Data and AI at commerceIQ, Payments at Bukuwarung (Fintech - YC20) and Search/Recommendation at BliBli (E-commerce - Unicorn). Seeking next exciting opportunity. Open to new tech stacks. If you want to work with me let's have a quick call : https://calendly.com/jay-joshi-k/workwithme-icebreaking-call

Location

Bengaluru

AI stack

antigravity

codex

cursor

Work experience

SDE 3 - Data + AI at CommerceIQ, Sep 2023 - Present · 3 yrs, Bengaluru, I own and scale CommerceIQ's core data orchestration platform—an Airflow-like system driving 300+ production workflows across 5 engineering teams on Databricks. Originally a basic SQL executor, I led its evolution into a robust, multi-language engine (supporting SQL, PySpark, Scala, and DBT) that is now the company standard for running data pipelines. Lately, I’ve been focusing on fixing bottlenecks in building AI agents and making the platform smarter: - Drastically Cut On-Call Stress: Built an internal RAG support assistant trained on 18 months of incident history and integrated it directly into our IDEs. This slashed mean time to resolution by 88% (from over 2 hours down to 15 minutes), giving the team massive breathing room. - Streamlined Dataset Onboarding: Designed a human-in-the-loop, agentic onboarding tool (via Claude Code) that compressed a painful, multi-day, 5-stage setup process into a single guided session—currently live and onboarding major retail customers. - Saved $200K+/Year in Infrastructure: Built an LLM-assisted SQL-to-Spark migration toolchain alongside workflow optimizations, instantly knocking over $200K/year off our Databricks compute bill. - Boosted Platform Stability & Scale: Anticipated a looming Databricks concurrency bottleneck to ship a multi-workspace routing architecture three weeks ahead of a major breach (tripling capacity headroom). Fixed recurring API throttling with Redis distributed rate limiting, and authored the company's distributed tracing standard. - Team Leadership: Serve as a design-review anchor and mentor for the team. Wrote the platform's support SLA and triage policies, which cut inbound ticket volumes by 50% year-over-year. Promoted to SDE 3 for driving platform architecture, proactive risk mitigation, and proven cost/reliability engineering.

Senior Software Engineer - Payments at BukuWarung, Apr 2023 - May 2023 · 2 mos, Bengaluru, Joined the team owning post-payment fulfillment for digital goods (bills, top-ups). Built the core order lifecycle and vendor integration layer, handling real-time delivery, async retries, and automatic vendor fallback routing to ensure reliable distributed system performance. - Emergency Payment Gateway Task Force: Selected for a high-pressure team to build an internal gateway from scratch in 30 days after our provider pulled out. Personally owned the low-level design, crafting an extensible data model for rapid bank onboarding and building disbursement monitoring with automated failovers. - Improved System Reliability & Testing: Refactored synchronous payment flows to async, eliminating timeout-related failures. Built a configurable mock service that replicated unpredictable bank sandbox behaviors, cutting QA cycle times in half. - Standardized Engineering Practices: Designed and drove adoption of a consistent, extensible API response framework used across all backend and frontend teams.

Software Engineer 2 - Payments at BukuWarung, Sep 2021 - Apr 2023 · 1 yr 8 mos, Bengaluru, Karnataka, India

Software Engineer - Search + Recommendations at Blibli, Jul 2019 - Sep 2021 · 2 yrs 3 mos, Bengaluru, Worked across search infrastructure, distributed systems, and applied ML for one of Indonesia's largest e-commerce platforms. - Led GCP & Kafka Migration: Containerized Java/Spring Boot services using Docker, integrated Vault and Consul, and moved product updates from REST calls to a Kafka-based, event-driven pipeline. Designed an exactly-once processing mechanism (Redis TTL + Postgres row-level locking) to keep data consistent across pods—completing the migration with zero rollbacks and beating the downtime schedule. - Fixed Solr Bottlenecks (10x Throughput): Diagnosed a major performance drag where flash sales and bulk stock refreshes triggered full-document rewrites on massive, 50-field Solr documents. Rewrote the system around Solr atomic partial updates, delivering a 10x improvement in update throughput under peak loads of 12,000 QPS. - Shipped Blibli’s First Live ML Inference Service: Independently built an NLP system to extract product attributes from search queries across a 20M+ catalog. Shipped the model-serving Python API using optimized TensorFlow (quantization/pruning), adding just ~40ms to p99 search latency while cutting clicks-to-"Add to Cart" by 20-25% in major categories.

Education

Bachelor's degree, Information Technology, Nirma University