Kamlesh Chhipa profile photo 1 on Tal

Kamlesh Chhipa

SDE 2 | Full Stack Developer | Python, Django, FastAPI, JavaScript, React.js, Next.js | AWS & GCP | Trino | Iceberg | Clickhouse | Kubernetes | Apache Spark | Bigquery | Freelancer

Bengaluru

5+ years of work experience

BrightEdge
Jun 20251 yr 3 mosPresent

Architected and scaled the core LLM processing infrastructure for AI Hyper Cube using Python, ClickHouse and MySQL, processing 5B+ tokens per month to run scoped LLM prompts that extracted intent, brand entities, sentiment, and citations from AI-generated prompts at production scale. Engineered a GPU-based LLM inference platform using vLLM, Vast.ai, and open-source models for large-scale entity extraction, reducing monthly AI processing costs by 62% ($40K → $15K). Built pre- and post-processing pipelines for AI Hyper Cube, processing 100M+ keyword SERPs using Python and BigQuery for prompt generation and Spark for post-processing with brand-level hash partitioning (MD5 modulo), enabling reliable analytics and UI consumption. Built a scalable Trino–Iceberg–ClickHouse data pipeline leveraging Python, GCS, Hive metadata, JavaScript UDFs, and Kubernetes. Created a custom memory-aware job scheduler and replaced the legacy BigQuery UDF process, achieving $18K–$20K in monthly savings while processing 1B+ events per month. Led the development of DCX Collector V2 leveraging Python, FastAPI, RabbitMQ, Redis, Docker, Argo Workflows, and BigQuery. Achieved a 4x faster processing speed, enabled seamless processing of 100M+ keywords per month across multiple vendor integrations. Implemented an automated AIO stitching system using Python, JavaScript, BigQuery UDFs, and Argo, mitigating Google num=100 deprecation and reducing DCX keyword collection costs ~60% across 25+ locales.
Hevo Data
Feb 20223 yrs 4 mosMay 2025

Engineered the Snowflake Partner Connect feature by exposing a secure POST API with a JWT fingerprint for automated destination creation, driving an additional 10-15 MQLs per month. Optimized the Signup API, reducing response time from 9 seconds to 2-2.5 seconds (≈75% decrease) and increasing signup conversion rates by 4-5%, enhancing the overall user experience. Led end-to-end development of a secure referral portal using JavaScript Promises to sync job submissions across the database, webhooks, and Lever API, increasing referral engagement by 40%. Migrated the Django website frontend to a Next.js app, incorporating TypeScript and Tailwind CSS, and Dockerized the app, achieving a 15% improvement in application load times. Implemented A/B testing on high-conversion pages (signup, demo) for a Django website using Jinja templates, HTML, SCSS, and JavaScript, resulting in a 20% increase in demo requests. Enhanced Core Web Vitals by deploying Varnish in-memory caching (improving good LCP URLs from 55.73% to 65.23%) and migrating to the CloudFront CDN, further raising good LCP URLs to 83.85%.
Hevo Data
Feb 20224 mosMay 2022

Bengaluru, Karnataka, India

• Developed Python scripts for automating the upload of Confluence documents to zendesk KBs, enhancing workflow efficiency for the customer experience team. • Designed and implemented a comprehensive dashboard that allowed the customer experience team to manage and monitor various automation scripts and operations, streamlining internal processes across the company.
Maxtap
Nov 20214 mosFeb 2022

Bangalore Rural, Karnataka, India

• Researched Technical Backend Architecture and used Python, Machine Learning and MySQL. Enhanced results from Google Video Intelligence API and Achieved ~90% Accuracy. • Created product UI/UX mockup front-end design on Adobe XD tool and Built a Chrome Extension from Tampermonkey Javascript code to integrate with 5+ OTT platforms such as Oho-Gujarati, Ullu, MX Player and Hoststar. • Reduced Project Costs by 10% by Creating an ETL pipeline using Javascript and Python to fetch data from 3+ Data Sources and store in the Firebase database and Google Cloud Bucket • Evaluated over 10+ Machine Learning APIs and analyzed the results to determine what is most suitable for product requirements.
DUIT
Jan 202110 mosOct 2021

New Delhi, Delhi, India

• Reduced project costs by 15% by utilizing a Python script to seed data for shops registered on Google using the Google Places API and storing the data in a Firebase database. • Reduced project costs by 10% by utilizing a Python script to seed data for shops registered on Google using the Google Places API and storing the data in a firebase database. • Improved mobile app engagement by 20% through UI design using Adobe XD and Flutter.
DUIT
Sep 20204 mosDec 2020

India

• Created ETL pipeline using JavaScript on Google Apps Script that fetched real-time data from a stock API and loaded in Firebase and Google Bigquery, resulting in a 20% reduction in project costs.
GitHub
@Kamlesh480
Medium
@kamleshchhipa480
LeetCode
@kamleshchhipa480
Personal portfoliokamlesh-chhipa.com

Kamlesh’s AI Stack

claude code
claude code
claude
claude
codex
codex
cursor
cursor
gemini
gemini
claudecode
claudecode

Education

jaypee institute of information technology

Bachelor's Degree, Computer Software Engineering

jaypee institute of information technology

20184 yrs2022

Join Kamlesh and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Kamlesh Chhipa on Tal

Headline

SDE 2 | Full Stack Developer | Python, Django, FastAPI, JavaScript, React.js, Next.js | AWS & GCP | Trino | Iceberg | Clickhouse | Kubernetes | Apache Spark | Bigquery | Freelancer

About

Backend Software Engineer (SDE-II) with 4+ years of experience building large-scale data platforms and LLM-driven systems for product and AI-first companies. I specialise in Python-based backend engineering, distributed data pipelines, and production AI infrastructure while remaining comfortable owning frontend and product-facing features end-to-end when needed. At BrightEdge, I work on AI Cube, where I architect and operate LLM processing infrastructure handling 5B+ tokens per month. My work spans the full lifecycle from SERP and AI Overview ingestion to LLM execution and analytics-ready outputs using Python, ClickHouse, MySQL, BigQuery, Spark, Trino, Iceberg, Kubernetes, and GCS. I focus on building systems that are deterministic, cost-efficient, and reliable at scale. I’ve delivered backend and data platforms with direct business impact, including: A Trino–Iceberg–ClickHouse pipeline that replaced legacy BigQuery UDF workflows, saving $18K–$20K monthly while processing 1B+ events per month. A fault-tolerant SERP ingestion system processing 100M+ keywords monthly with 4× faster throughput. An AI Overview stitching pipeline that reduced data collection costs by ~60% across 25+ locales. A GSC BigQuery export pipeline adopted by 50+ enterprise customers (including DIRECTV and LinkedIn), improving data accuracy by ~10% Previously at Hevo Data, I worked as a Full Stack Developer, owning backend APIs and frontend experiences end-to-end. I built secure integrations like Snowflake Partner Connect (driving 10–15 MQLs/month), optimized critical APIs (~75% latency reduction, +4–5% signup conversion), and led product-facing initiatives including Next.js migrations, A/B experimentation (+20% demo requests), referral systems (+40% engagement), and Core Web Vitals improvements (LCP improved to 83.85%). I enjoy building backend-first systems that scale, while staying hands-on with frontend and product engineering when it drives impact. Open to SDE-2 / SDE-3 Backend, Platform, or Full-Stack roles in product-led, AI-driven teams.

Location

Bengaluru

AI stack

claude code

claude

codex

cursor

gemini

claudecode

Work experience

Software Engineer (SDE 2) at BrightEdge, Jun 2025 - Present · 1 yr 3 mos, Architected and scaled the core LLM processing infrastructure for AI Hyper Cube using Python, ClickHouse and MySQL, processing 5B+ tokens per month to run scoped LLM prompts that extracted intent, brand entities, sentiment, and citations from AI-generated prompts at production scale. Engineered a GPU-based LLM inference platform using vLLM, Vast.ai, and open-source models for large-scale entity extraction, reducing monthly AI processing costs by 62% ($40K → $15K). Built pre- and post-processing pipelines for AI Hyper Cube, processing 100M+ keyword SERPs using Python and BigQuery for prompt generation and Spark for post-processing with brand-level hash partitioning (MD5 modulo), enabling reliable analytics and UI consumption. Built a scalable Trino–Iceberg–ClickHouse data pipeline leveraging Python, GCS, Hive metadata, JavaScript UDFs, and Kubernetes. Created a custom memory-aware job scheduler and replaced the legacy BigQuery UDF process, achieving $18K–$20K in monthly savings while processing 1B+ events per month. Led the development of DCX Collector V2 leveraging Python, FastAPI, RabbitMQ, Redis, Docker, Argo Workflows, and BigQuery. Achieved a 4x faster processing speed, enabled seamless processing of 100M+ keywords per month across multiple vendor integrations. Implemented an automated AIO stitching system using Python, JavaScript, BigQuery UDFs, and Argo, mitigating Google num=100 deprecation and reducing DCX keyword collection costs ~60% across 25+ locales.

Full Stack Engineer at Hevo Data, Feb 2022 - May 2025 · 3 yrs 4 mos, Engineered the Snowflake Partner Connect feature by exposing a secure POST API with a JWT fingerprint for automated destination creation, driving an additional 10-15 MQLs per month. Optimized the Signup API, reducing response time from 9 seconds to 2-2.5 seconds (≈75% decrease) and increasing signup conversion rates by 4-5%, enhancing the overall user experience. Led end-to-end development of a secure referral portal using JavaScript Promises to sync job submissions across the database, webhooks, and Lever API, increasing referral engagement by 40%. Migrated the Django website frontend to a Next.js app, incorporating TypeScript and Tailwind CSS, and Dockerized the app, achieving a 15% improvement in application load times. Implemented A/B testing on high-conversion pages (signup, demo) for a Django website using Jinja templates, HTML, SCSS, and JavaScript, resulting in a 20% increase in demo requests. Enhanced Core Web Vitals by deploying Varnish in-memory caching (improving good LCP URLs from 55.73% to 65.23%) and migrating to the CloudFront CDN, further raising good LCP URLs to 83.85%.

Tech Intern | Python, MySQL, REST API, ETL, Django at Hevo Data, Feb 2022 - May 2022 · 4 mos, Bengaluru, Karnataka, India, • Developed Python scripts for automating the upload of Confluence documents to zendesk KBs, enhancing workflow efficiency for the customer experience team. • Designed and implemented a comprehensive dashboard that allowed the customer experience team to manage and monitor various automation scripts and operations, streamlining internal processes across the company.

Software Engineer | Python, MySQL, Machine Learning, OTT Integration at Maxtap, Nov 2021 - Feb 2022 · 4 mos, Bangalore Rural, Karnataka, India, • Researched Technical Backend Architecture and used Python, Machine Learning and MySQL. Enhanced results from Google Video Intelligence API and Achieved ~90% Accuracy. • Created product UI/UX mockup front-end design on Adobe XD tool and Built a Chrome Extension from Tampermonkey Javascript code to integrate with 5+ OTT platforms such as Oho-Gujarati, Ullu, MX Player and Hoststar. • Reduced Project Costs by 10% by Creating an ETL pipeline using Javascript and Python to fetch data from 3+ Data Sources and store in the Firebase database and Google Cloud Bucket • Evaluated over 10+ Machine Learning APIs and analyzed the results to determine what is most suitable for product requirements.

Software Engineer | Python, Google Places API, Firebase, ETL Pipeline at DUIT, Jan 2021 - Oct 2021 · 10 mos, New Delhi, Delhi, India, • Reduced project costs by 15% by utilizing a Python script to seed data for shops registered on Google using the Google Places API and storing the data in a Firebase database. • Reduced project costs by 10% by utilizing a Python script to seed data for shops registered on Google using the Google Places API and storing the data in a firebase database. • Improved mobile app engagement by 20% through UI design using Adobe XD and Flutter.

Software Engineer Intern | Javascript, Firebase, Google BigQuery, API Integration at DUIT, Sep 2020 - Dec 2020 · 4 mos, India, • Created ETL pipeline using JavaScript on Google Apps Script that fetched real-time data from a stock API and loaded in Firebase and Google Bigquery, resulting in a 20% reduction in project costs.

Education

Bachelor's Degree, Computer Software Engineering, jaypee institute of information technology

Public links

github: https://github.com/Kamlesh480

medium: https://medium.com/@kamleshchhipa480

leetcode: https://leetcode.com/kamleshchhipa480

kamlesh-chhipa.com: https://kamlesh-chhipa.com/