Dheer Varyani profile photo 1 on Tal

Dheer Varyani

Software Engineer II @ Swiggy | Backend Systems · Distributed Systems · Java & Go | Ex-Meesho,Cropin · IIIT Bangalore

Bengaluru

3+ years of work experience

Swiggy
Jul 20262 mosPresent

Bengaluru

Meesho
Dec 20258 mosJul 2026

RTO Prediction & Carrier Allocation Engine — Redesigned MongoDB data model from 1 billion flat documents to 1 million hierarchical (1000× reduction); built a custom 64-bit packed binary Redis encoder storing 6 fields in 8 bytes (vs ~100-byte JSON) — 12.5× memory reduction, saving 69 GB across 21M cache keys. RTO Data Pipeline — Designed Kafka consumer with main → retry → DLQ pattern, batch MongoDB upserts, and async Redis warm-up; 3-tier cache-first fetch (Redis → MongoDB) with 30ms SLA; integrated RTO rate and SubOCF as enrichment signals into carrier ranking and filtering. Mentorship & On-Call — Mentor 4 SDE-1 engineers on distributed systems design and caching patterns; lead on-call response for critical production issues including zone cron master/slave misconfiguration (order loss), COD serviceability regression, and ShipMatrix CPU starvation.
Meesho
Sep 20241 yr 4 mosDec 2025

Bengaluru

Pin-to-Pin Zone Mapping — Overhauled PDD (Promised Date of Delivery) computation to leverage granular pin-level data — ₹0.40 CPDO improvement per order. Scalable Cron Pipeline — Processed 22.5 billion records across 25K × 25K pincode pairs, 6 carriers, 3 account types, 2 payment modes; scaled zone definitions from 25 to 25,000; drove end-to-end ownership including sale-event scaling based on traffic prediction. Cache Optimisation & Capacity Migration — Redesigned cache key structure, improving hit rate from 40% to 80%+ and cutting monthly Redis spend by $1,000; migrated 23.6B capacity keys via dual-write strategy with zero downtime. ShipMatrix — Candidate Generation Service (Go) — End-to-end ownership of production system for candidate generation, promise computation, and capacity management; architected MongoDB + Redis (protobuf) data flow with 7+ write paths and resilient concurrency design.
Cropin
Jun 20231 yr 5 mosOct 2024

Bengaluru

Revamped 7 production APIs scaling data-handling capacity 10× (3K → 30K records) and reducing response time from 30s to under 2s; collaborated cross-functionally with product and data teams on technical design and delivery. Built custom attribute system supporting 60K farmers/assets per tenant with sub-2s filtering and full-text search using ElasticSearch, Redis, a
Cropin
Jan 20236 mosJun 2023

On-site

Contributed to backend feature development on Spring Boot services; converted to full-time SDE-1 in June 2023.
GitHub
@dheer009

Dheer’s AI Stack

cursor
cursor
opencode
opencode

Education

International Institute of Information Technology Bangalore

Master of Technology - MTech, Computer Science

International Institute of Information Technology Bangalore

20212 yrs2023

Bachelor of Technology - BTech, Computer Engineering

SILVER OAK COLLEGE OF ENGG., & TECH., AHMEDABAD 077

20164 yrs2020

Join Dheer and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Dheer Varyani on Tal

Headline

Software Engineer II @ Swiggy | Backend Systems · Distributed Systems · Java & Go | Ex-Meesho,Cropin · IIIT Bangalore

About

Senior Backend Engineer with 3+ years building high-throughput distributed systems in Java and Go. At Meesho, I work on the logistics platform that powers 80 lakh+ daily shipments — designing data models, caching architectures, and Kafka pipelines that serve sub-30ms SLAs across billions of records. Recent work includes redesigning a 1-billion-document MongoDB collection down to 1 million (1000× reduction), and building a custom 64-bit packed binary Redis encoder that cut cache memory by 12.5× — saving 69 GB across 21M keys. Previously at Cropin, I revamped 7 production APIs to scale 10× and built a custom attribute system for 60K farmers per tenant using ElasticSearch, Redis, and Kafka. Interests: data modelling, caching strategy, system design, API design, and the messy real-world tradeoffs in scaling backend systems. M.Tech in CS, IIIT Bangalore (GATE CS 2021 — AIR 426, 99.58 percentile). Tech: Java · Go · Spring Boot · MongoDB · Redis · Kafka · Protobuf · PostgreSQL · ElasticSearch · AWS · Docker

Location

Bengaluru

AI stack

cursor

opencode

Work experience

software development engineer 2 at Swiggy, Jul 2026 - Present · 2 mos, Bengaluru

Software Engineer-2 at Meesho, Dec 2025 - Jul 2026 · 8 mos, RTO Prediction & Carrier Allocation Engine — Redesigned MongoDB data model from 1 billion flat documents to 1 million hierarchical (1000× reduction); built a custom 64-bit packed binary Redis encoder storing 6 fields in 8 bytes (vs ~100-byte JSON) — 12.5× memory reduction, saving 69 GB across 21M cache keys. RTO Data Pipeline — Designed Kafka consumer with main → retry → DLQ pattern, batch MongoDB upserts, and async Redis warm-up; 3-tier cache-first fetch (Redis → MongoDB) with 30ms SLA; integrated RTO rate and SubOCF as enrichment signals into carrier ranking and filtering. Mentorship & On-Call — Mentor 4 SDE-1 engineers on distributed systems design and caching patterns; lead on-call response for critical production issues including zone cron master/slave misconfiguration (order loss), COD serviceability regression, and ShipMatrix CPU starvation.

Software Engineer-1 at Meesho, Sep 2024 - Dec 2025 · 1 yr 4 mos, Bengaluru, Pin-to-Pin Zone Mapping — Overhauled PDD (Promised Date of Delivery) computation to leverage granular pin-level data — ₹0.40 CPDO improvement per order. Scalable Cron Pipeline — Processed 22.5 billion records across 25K × 25K pincode pairs, 6 carriers, 3 account types, 2 payment modes; scaled zone definitions from 25 to 25,000; drove end-to-end ownership including sale-event scaling based on traffic prediction. Cache Optimisation & Capacity Migration — Redesigned cache key structure, improving hit rate from 40% to 80%+ and cutting monthly Redis spend by $1,000; migrated 23.6B capacity keys via dual-write strategy with zero downtime. ShipMatrix — Candidate Generation Service (Go) — End-to-end ownership of production system for candidate generation, promise computation, and capacity management; architected MongoDB + Redis (protobuf) data flow with 7+ write paths and resilient concurrency design.

Software Engineer at Cropin, Jun 2023 - Oct 2024 · 1 yr 5 mos, Bengaluru, Revamped 7 production APIs scaling data-handling capacity 10× (3K → 30K records) and reducing response time from 30s to under 2s; collaborated cross-functionally with product and data teams on technical design and delivery. Built custom attribute system supporting 60K farmers/assets per tenant with sub-2s filtering and full-text search using ElasticSearch, Redis, a

Intern at Cropin, Jan 2023 - Jun 2023 · 6 mos, On-site, Contributed to backend feature development on Spring Boot services; converted to full-time SDE-1 in June 2023.

Education

Master of Technology - MTech, Computer Science, International Institute of Information Technology Bangalore

Bachelor of Technology - BTech, Computer Engineering, SILVER OAK COLLEGE OF ENGG., & TECH., AHMEDABAD 077

Public links

github: https://github.com/dheer009