Started Web3 seriously this year. Here’s the journey so far
→ Won $76,000+ in grants & hackathons
→ Built & shipped Solana trading and MEV infrastructure
(launchpad, SDKs, bots, indexers)
→ Open sourced multiple MEV tools & trackers
→ Earned 200+ GitHub stars across OSS projects
→ Did freelance trading/MEV bot work for real users
→ Lost a lot of money trading perps learned risk the hard way
→ Turned those lessons into better infra & risk aware systems
→ Landed a Web3 infra role, building core execution and data infrastructure
Solana didn’t just give me wins it gave me hard lessons, leverage, and belief.
No IIT. No MIT.
No fancy resume.
Just consistent shipping, OSS, and building in public.
Thank you @solana for an unreal 2025.
2026 will be even bigger.
Your turn what’s your journey?
2025 was about turning momentum into outcomes. Here’s the journey so far:
→ Scaled one of my businesses to $1.8M
→ Explored 3 continents, 8 countries, and multiple cities
→ Invested deeply in tech and built real products
→ Contributed to YC-backed open source projects
→ Landed two jobs combined worth $100K+ USD
→ Worked closely with strong growth and marketing teams in the creator economy
→ Started talking to someone invaluable who now holds deep importance in my life
→ Gifted a house to my parents
→ Applied to YC
→ Started gaming again
No IIT. No MIT. No fancy job.
Just relentless execution and consistent output.
2026 is about scale.
What's your story?
If you’ve ever parsed a Raydium swap and had to filter Program TokenkegQ… invoke [2] noise out of log_messages(), I just shipped direct_log_messages() in yellowstone-vixen.
On mainnet, a SwapV2 call’s logs drop from 11 lines to 5 no CPI chatter, just your program’s own logs.
@triton_one@solana_devs@Raydium@readylayerone@linuskendall
Pinocchio-native CPMM on Solana init, deposit, swap, withdraw.
Mollusk run against the deployed .so; math is byte-exact to x · y = k at every step:
init: 22,851 CU
deposit: 22,458 CU
swap: 16,542 CU
withdraw: 22,719 CU
Devnet:4rP1EFZ8BPBprGPdcDtduidTPjwzXxLmGqvrxLmCiTQX
NOTE :Not audited, not mainnet, repo soon
Got my PR merged into yellowstone-vixen.
Jetstream now:
-> Forwards real block entries (no more empty but counted blocks)
-> Exits cleanly on SIGINT
This makes it much easier to build reliable block stream consumers and monitoring on top of it. :)
Solana blockhashes expire after ~60s. For a trading bot that’s too long you want the tx to die in ~10s if it doesn’t land, not sit there executing on stale prices.
Shipped a PR to solana-blockhash-cache fixing dead commitment config + a backfiller stall, and documented the older-blockhash trick that lets you pick the expiry window.
Better fills don't happen by accident.
Partnering with @DFlow means every Solana trade on Boba routes through powerful trading infrastructure.
Our users get the best, because we build with the best.
@Circular_fi yes sir! still its more then 2 million 🫡
Gross USDC profit - $2,115,737.65
Total SOL spent (fees+tips) - 359.10 SOL = $31,960
NET arb profit - $2,083,778
$2,993,502 drained from a single @solana pool in 150 seconds
1 whale. 6 bots. 9 transactions. 1 misconfigured @MeteoraAG pool that sat for 10 months
Many of you asked how. Here's the full onchain breakdown 🧵👇
Was looking into arb bots on @solana found one of the craziest arbs I’ve seen
$0.39 → $1.1M in <1s
Two bots rinsed the same stale @MeteoraAG pool in consecutive blocks
Bot 1: $0.16 → $403K
Bot 2: $0.23 → $696K
~800ms apart
Same pool. Different bots
Was looking into arb bots on @solana found one of the craziest arbs I’ve seen
$0.39 → $1.1M in <1s
Two bots rinsed the same stale @MeteoraAG pool in consecutive blocks
Bot 1: $0.16 → $403K
Bot 2: $0.23 → $696K
~800ms apart
Same pool. Different bots
@aabuzarghaffari@solana@MeteoraAG It was mainly due to a price mismatch between two pools the Meteora pool was priced significantly higher than others. Bots exploited this by buying from cheaper pools and selling through Meteora DLMM, ultimately draining USDC liquidity from the pool
@Circular_fi@solana@MeteoraAG Yeah thanks, but total wipeout was around $2.1M, driven by 9 major arb transactions in two consecutive blocks. It all kicked off with a whale dumping ~40% of the total token supply through Jupiter
Logs just disappear, no error. Took a while to track down.
The log classifier was treating " invoke [" and " success" substrings inside Program log: payloads as real CPI events, so any program that logged user-controlled strings could truncate its own log_range.
Fixed in #249. Same PR adds direct_log_messages() — CPI-aware iterator, only yields logs at your program’s invocation depth. If you’re parsing @Raydium / @pumpdotfun / @orca_so / @MeteoraAG swaps, use that one.
merged into anza-xyz/jetstreamer again 🫡
thanks @Sam0xSol — wasn't much this time, but shipping is shipping.
Solana protocol-level Rust, every week.
https://t.co/JbkAYGQR4q