abhishek kumar profile photo 1 on Tal

abhishek kumar

Head of Engineering | Leading AI initiative | Ex Meta, Agoda, Tesco

Bengaluru, Karnataka

9+ years of work experience

Stealth Mode Startup
Jul 20251 yr 1 moPresent

Remote

-> Built the engineering team from zero to 10. Set up sprint cadences, defined how we work, and drove hiring alongside the founding PM. -> Migrated payment infrastructure from Stripe to Flowt/Paynetworkx with zero downtime. Renegotiated CC processing rates through Paynetworkx, significantly reducing transaction costs. These savings compound meaningfully at scale. Built the payment state machines and webhook handlers to make it reliable. -> Architected an agentic AI copilot from scratch that converts natural language queries into real system actions across the marketplace. Built a tool-calling framework with RBAC so customers can check request/booking status, while state-override and provider-management tools are locked to platform admins only. -> Designed the end-to-end booking lifecycle as a state machine with 13+ states covering request, proposal, payment, and cancellation flows across customer and provider actors, with handling for expired requests, withdrawn proposals, and backend-initiated cancellations. -> Set up observability from scratch using CloudWatch and X-Ray across all microservices. Kept it cost-effective on AWS-native tooling rather than going with expensive third-party APM.
Tesco
Jun 20232 yrs 2 mosJul 2025

-> Managed 8 engineers across backend and platform. Owned sprint planning, roadmaps, and worked closely with product, design, and ops leads. ->Led the build of an enterprise CMS backend that brought content creation latency from seconds down to <100ms. Editor productivity went up 60%. -> Redesigned the service architecture into a single multi-module runtime, which cut memory-related crashes and reduced production restarts by 99%. -> Built a rule-driven personalization system with a Drools-based decision engine, serving targeted content to 10M+ users. -> Re-architected the buylist API: P99 dropped from 250ms to 150ms at 100K+ RPM. -> Drove RSocket migration across 15+ microservices. 25% average response time improvement. -> Set up code review standards, on-call rotations, and incident response processes. Production incidents dropped 30%.
Meta
Jan 20221 yr 5 mosMay 2023

Greater London, England, United Kingdom

Worked on commerce infrastructure — catalog ingestion, data pipelines, and the TAO distributed data store. → Redesigned a high-contention TAO update pipeline using async write coalescing. Reduced optimistic lock exceptions by 99.5% at billion-scale → Fixed recurring batch pipeline failures by tuning execution timing and retry windows → Built a rate-limiting system for catalog ingestion based on supplier identification → Architected an experimentation framework for catalog data analysis processing 1B+ items daily
Embibe
Nov 20192 yrs 1 moNov 2021

Bangalore Urban, Karnataka, India

-> Led a team of 12 engineers at Embibe, an edtech startup. Did a lot of hiring — ran technical interviews and helped grow the engineering org during a period of rapid scaling. -> Biggest project was migrating User Auth off a Ruby on Rails monolith to Java microservices, which cut infra costs by 40% and made response times 50% faster. Also architected UserHome and Timeline services from scratch using event-driven architecture. -> Built the "ASK" platform — basically Embibe's version of Stack Overflow — serving 50K+ daily active users with sub-200ms response times. Towards the end, also handled the blog/CMS migration off legacy WordPress. Stack: Java, Spring Boot, Spring WebFlux, Elasticsearch, ScyllaDB/Cassandra, Spark, Docker, Redis, MongoDB, Kubernetes
Agoda
Aug 20163 yrs 3 mosOct 2019

Bangkok, Bangkok City, Thailand

Backend and data engineering on Agoda's pricing and infrastructure team in Bangkok. - Pricing system serving 500K RPS across 5 global data centers and 200+ servers. Optimized the calculation engine — brought response times from 600ms down to 450ms - ETL pipeline optimization: cut runtime from 4.5 hours to 3 hours. Separately built data pipelines moving 10TB+ daily from RDS to S3 for analytics - Kafka infrastructure — created Terraform/Packer IaC templates for automated cluster deployment, added OAuth 2.0 + RBAC - Built an A/B testing analysis tool (React + Python) that ended up being used by 50+ PMs across the org Stack: Java, Scala, Python, Akka, Spark, Kafka, Terraform, AWS

abhishek’s AI Stack

claude code
claude code
claude
claude
langchain
langchain
langgraph
langgraph

Education

Indian Institute of Technology, Patna

Bachelor of Technology, Computer Engineering

Indian Institute of Technology, Patna

20094 yrs2013
kendriya vidyalaya

10+2

kendriya vidyalaya

20062 yrs2008

Join abhishek and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

abhishek kumar on Tal

Headline

Head of Engineering | Leading AI initiative | Ex Meta, Agoda, Tesco

About

Results-oriented Engineering Leader with over 13 years of experience building resilient distributed systems and scaling high-performing teams at Meta, Agoda, and Tesco. Expertise spans enterprise-grade infrastructure, microservices, and AI-driven product development. Proven track record in driving architectural migrations, optimizing performance, and fostering operational excellence to deliver measurable business impact at scale.

Location

Bengaluru, Karnataka

AI stack

claude code

claude

langchain

langgraph

Work experience

Director of Engineering at Stealth Mode Startup, Jul 2025 - Present · 1 yr 1 mo, Remote, -> Built the engineering team from zero to 10. Set up sprint cadences, defined how we work, and drove hiring alongside the founding PM. -> Migrated payment infrastructure from Stripe to Flowt/Paynetworkx with zero downtime. Renegotiated CC processing rates through Paynetworkx, significantly reducing transaction costs. These savings compound meaningfully at scale. Built the payment state machines and webhook handlers to make it reliable. -> Architected an agentic AI copilot from scratch that converts natural language queries into real system actions across the marketplace. Built a tool-calling framework with RBAC so customers can check request/booking status, while state-override and provider-management tools are locked to platform admins only. -> Designed the end-to-end booking lifecycle as a state machine with 13+ states covering request, proposal, payment, and cancellation flows across customer and provider actors, with handling for expired requests, withdrawn proposals, and backend-initiated cancellations. -> Set up observability from scratch using CloudWatch and X-Ray across all microservices. Kept it cost-effective on AWS-native tooling rather than going with expensive third-party APM.

Engineering Manager at Tesco, Jun 2023 - Jul 2025 · 2 yrs 2 mos, -> Managed 8 engineers across backend and platform. Owned sprint planning, roadmaps, and worked closely with product, design, and ops leads. ->Led the build of an enterprise CMS backend that brought content creation latency from seconds down to <100ms. Editor productivity went up 60%. -> Redesigned the service architecture into a single multi-module runtime, which cut memory-related crashes and reduced production restarts by 99%. -> Built a rule-driven personalization system with a Drools-based decision engine, serving targeted content to 10M+ users. -> Re-architected the buylist API: P99 dropped from 250ms to 150ms at 100K+ RPM. -> Drove RSocket migration across 15+ microservices. 25% average response time improvement. -> Set up code review standards, on-call rotations, and incident response processes. Production incidents dropped 30%.

Senior Software Engineer at Meta, Jan 2022 - May 2023 · 1 yr 5 mos, Greater London, England, United Kingdom, Worked on commerce infrastructure — catalog ingestion, data pipelines, and the TAO distributed data store. → Redesigned a high-contention TAO update pipeline using async write coalescing. Reduced optimistic lock exceptions by 99.5% at billion-scale → Fixed recurring batch pipeline failures by tuning execution timing and retry windows → Built a rate-limiting system for catalog ingestion based on supplier identification → Architected an experimentation framework for catalog data analysis processing 1B+ items daily

Lead Engineer at Embibe, Nov 2019 - Nov 2021 · 2 yrs 1 mo, Bangalore Urban, Karnataka, India, -> Led a team of 12 engineers at Embibe, an edtech startup. Did a lot of hiring — ran technical interviews and helped grow the engineering org during a period of rapid scaling. -> Biggest project was migrating User Auth off a Ruby on Rails monolith to Java microservices, which cut infra costs by 40% and made response times 50% faster. Also architected UserHome and Timeline services from scratch using event-driven architecture. -> Built the "ASK" platform — basically Embibe's version of Stack Overflow — serving 50K+ daily active users with sub-200ms response times. Towards the end, also handled the blog/CMS migration off legacy WordPress. Stack: Java, Spring Boot, Spring WebFlux, Elasticsearch, ScyllaDB/Cassandra, Spark, Docker, Redis, MongoDB, Kubernetes

Senior Software Engineer at Agoda, Aug 2016 - Oct 2019 · 3 yrs 3 mos, Bangkok, Bangkok City, Thailand, Backend and data engineering on Agoda's pricing and infrastructure team in Bangkok. - Pricing system serving 500K RPS across 5 global data centers and 200+ servers. Optimized the calculation engine — brought response times from 600ms down to 450ms - ETL pipeline optimization: cut runtime from 4.5 hours to 3 hours. Separately built data pipelines moving 10TB+ daily from RDS to S3 for analytics - Kafka infrastructure — created Terraform/Packer IaC templates for automated cluster deployment, added OAuth 2.0 + RBAC - Built an A/B testing analysis tool (React + Python) that ended up being used by 50+ PMs across the org Stack: Java, Scala, Python, Akka, Spark, Kafka, Terraform, AWS

Education

Bachelor of Technology, Computer Engineering, Indian Institute of Technology, Patna

10+2, kendriya vidyalaya