Open to opportunities

Keval Dave

@kevaldave

Full-stack engineer building scalable analytics and AI platforms that measurably improve product decisions and performance.

India

Message

What I'm looking for

I’m looking for a team where I can ship end-to-end analytics and AI features—focused on clean architecture, strong developer experience, and measurable product impact, backed by observability, reliable APIs, and fast iteration.

I’m a full-stack engineer with 5+ years of experience building scalable systems across fintech, e-commerce, logistics, and enterprise domains, with a strong focus on clean architecture and developer experience.

In my current role, I’m building a strategic analytics platform for campaign visualisation, scenario modelling, and marketing spend decisions using Python (FastAPI) and React, including production RAG and agentic AI workflows for natural-language querying and autonomous multi-step analysis.

I also instrumented AI systems for end-to-end observability using Langfuse—tracking LLM usage, latency, token costs, and surfacing production issues—while designing AI agents with Google ADK + Vertex AI for tool-use across internal analytics data sources.

Previously, I delivered major reliability and performance wins at scale: migrating affiliate attribution to Node.js/TypeScript microservices on AWS DynamoDB, re-architecting login services with Kafka event tracking and Redis sessions, improving user experience via profiling and Core Web Vitals fixes, and automating serverless data pipelines into AWS Redshift with full CI/CD.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer, Analytics AI

Current

Colgate-Palmolive

Oct 2025 - Present (10 months)

Built a strategic analytics platform for campaign visualization, scenario modeling, and marketing spend decisions using Python (FastAPI) and React (Plotly). Integrated RAG and agentic AI workflows (Google ADK, Vertex AI) with Langfuse observability and implemented user auth with PostgreSQL and Redis-backed agent session state.

Python fastAPI React Plotly Agentic AI Google ADK Vertex AI Langfuse LLM Observability PostgreSQL Redis Multi Step Reasoning Tool Use

Software Development Engineer II

Shaadi.com — People Interactive India Pvt Ltd

Aug 2024 - Oct 2025 (1 year 2 months)

Migrated affiliate attribution and tracking to a Node.js/TypeScript microservice on AWS DynamoDB, and re-architected login using Kafka event tracking and Redis sessions. Worked on React/Next.js performance (SEO/SSR/SSG, Web Vitals), serverless Python data pipelines to Redshift, and improved developer/build efficiency with SWC, maintaining Jest/Cypress coverage across repos.

Technical Consultant (Contractual)

Softpixel Solutions → Quantiphi Analytics

Aug 2023 - Aug 2024 (1 year)

Contracted to build client platforms and embedded purchase flows using Next.js/TypeScript and React Native SDK integration with Google Pay and HDFC PayZapp. Designed a polling-based payment reconciliation system with a backend using Python Flask and PostgreSQL, including mentoring on React and API integration.

Next.js TypeScript React Native Google Pay PayZapp Python Flask PostgreSQL Payment Reconciliation API Integration Mentor Polling Systems CI CD

Software Dev Engineer (Full Stack)

Terribly Tiny Tales Pvt Ltd

Jun 2021 - Dec 2022 (1 year 6 months)

Built a Next.js frontend and Python FastAPI backend for Fambase and ttt.community using SSR and SSG for SEO-optimized rendering and fast initial load times. Developed backend services including a Zoom meeting management workflow, a WhatsApp Business API microservice on AWS Lambda, and a referral & cashback system.

Next.js Python fastAPI SEO AWS Lambda WhatsApp Business API Microservices Zoom Management Referrals Cashback

Education

Degrees, certifications, and relevant coursework

K.J. Somaiya Institute of Engineering & IT

Bachelor of Engineering, Electronics & Telecommunications

Grade: 7.27/10

B.E. in Electronics & Telecommunications at K.J. Somaiya Institute of Engineering & IT. The provided text lists conflicting time ranges (2017–2021 and Jun 2021–Dec 2022) and shows a grade of 7.27/10.