I’m a Senior Staff Software Engineer at Sharechat. I love reading and writing about distributed systems, performance issues, and high-load solutions.
London, UK
Joined December 2012
- Tweets66
- Following228
- Followers40
- Likes501
Pinned Tweet
I just finished my talk at #current2025 . The recording will be available soon. But you can find my slides already here docs.google.com/presentation…
Thanks @confluentinc for the opportunity to talk at the Current!
Andrei Manakov retweeted
ShareChat's Node.js + Redis deduplication system could not scale cost-effectively. At @P99CONF, @AndreyManakov and Sanyam Gupta explain how they migrated to Apache Flink and ultimately achieved a 7x cost reduction. Watch for free here > ow.ly/gCm050XowPw
#ScyllaDB #P99CONF
Andrei Manakov retweeted
@ShareChatapp costs spiked as they scaled to 300 million users who generated billions of daily interactions. @AndreyManakov shares how moving to Flink, #ScyllaDB, & @Redpandadata achieved a 90% cost reduction while lowering P99 latency from 40 ms to 8 ms. scylladb.com/2026/03/02/how-…
Andrei Manakov retweeted
ShareChat operates dozens of #ScyllaDB clusters that power multiple products. In 15 minutes, @AndreyManakov shares their data-driven cost optimization process and how they reduced infrastructure costs without sacrificing performance. Join for free here > ow.ly/TNA650YskX4
Andrei Manakov retweeted
@ShareChatapp costs spiked as they scaled to 300 million users who generated billions of daily interactions. @AndreyManakov shares how moving to Flink, #ScyllaDB, & @Redpandadata achieved a 90% cost reduction while lowering P99 latency from 40 ms to 8 ms. scylladb.com/2026/03/02/how-…
I received a speaker swag from #scylladb it looks so cool, I specially love this scylla boy😍 Thanks @ScyllaDB !
Join me on Monster Scale summit, I will be talking about capacity analysis and best practices we have learned by running dozens of ScyllaDBs.
scylladb.com/monster-scale-s…
Andrei Manakov retweeted
@ShareChatapp's @AndreyManakov is a trusted voice when it comes to solving scaling challenges. During Monster Scale Summit, Andrei will share how they built a data-driven cost optimization process & reduced costs without sacrificing performance. ow.ly/RW6r50Y1f5N
#ScyllaDB
Andrei Manakov retweeted
@ShareChatapp's @AndreyManakov is a trusted voice when it comes to solving scaling challenges. During Monster Scale Summit, Andrei will share how they built a data-driven cost optimization process & reduced costs without sacrificing performance. ow.ly/RW6r50Y1f5N
#ScyllaDB
Andrei Manakov retweeted
To effectively analyze #ScyllaDB cluster capacity, there are three key recommendations to focus on.
ShareChat's @AndreyManakov shares monitoring tips that can help reduce cluster size 2-5X without compromising latency > ow.ly/C2Pt50XR3Bq
Andrei Manakov retweeted
ShareChat's Node.js + Redis deduplication system could not scale cost-effectively. At @P99CONF, @AndreyManakov and Sanyam Gupta explain how they migrated to Apache Flink and ultimately achieved a 7x cost reduction. Watch for free here > ow.ly/gCm050XowPw
#ScyllaDB #P99CONF
Recording of our P99 talk is already available
ShareChat's Node.js + Redis deduplication system could not scale cost-effectively. At @P99CONF, @AndreyManakov and Sanyam Gupta explain how they migrated to Apache Flink and ultimately achieved a 7x cost reduction. Watch for free here > ow.ly/gCm050XowPw
#ScyllaDB #P99CONF
It is a review on “On Writing Well” by William Zinsser - a classic guide to good writing. The book explains key principles: clarity, simplicity, and more. And you know what? Most of these principles are well-known to experienced developers andection.substack.com/p/cle…
Join Sanyam and me in P99. We are having the talk right now
ShareChat processes over 200K requests per second, but their Node.js + Redis deduplication system could not scale cost-effectively. In 10 minutes, @AndreyManakov & Sanyam Gupta discuss their Apache Flink migration & 7X cost reduction.
ow.ly/azUs50XgRrl
#ScyllaDB #P99CONF
Join me on next P99, I will be talking about real-time deduplication at scale
Why attend @P99CONF? In addition to speakers from Microsoft, Netflix, AWS, Meta, Uber, and Bloomberg, you don't want to miss these highly technical sessions on October 22-23 from @SivukhinN, @AndreyManakov, and many more. ow.ly/nOAX50WPGAO
#ScyllaDB #P99CONF
Andrei Manakov retweeted
Not quite 🙂
ShareChat & Moj together serve more than 20 crore monthly active users. Even third-party estimates like data.ai (which typically undercount Tier 2/3 traffic) put us at 12 crore MAUs.
Over the last two years, we have built one of the most cost-efficient recommender systems in the industry, enabling us to turn cashflow positive in Feb 2025. We have also been working on our monetization and are currently at a ₹1,000 crore annual revenue run rate.
We are now doubling down on engaging new content formats, especially micro-dramas, which already have scaled to 5 crore monthly active viewers on our platforms.
Far from “crashed”, ShareChat & Moj today stand strong with 20+ crore monthly active users, ₹1,000+ crore revenue, profitable and growing.
Jai Hind! 🇮🇳
#jetbrains is very fast! Last Junie version uses GPT-5 by default, which was released just yesterday!
Andrei Manakov retweeted
Excellent article from Brendan Gregg outlining what a Performance Engineering Team does, and when a company might consider hiring one:
brendangregg.com/blog/2025-0…
There is a really nice and short article about bloom filters llimllib.github.io/bloomfilt…
recording - current.confluent.io/post-co…