I build and scale real fintech systems.
₹10,000+ Cr payment volume (Cards/UPI/Credit)
Scaled from <10 TPS → 250+ TPS
99.99% uptime in production
Lately focused on AI:
+25% dev productivity in 3 months
Delivery cycle: 8 weeks → 4
Sharing what actually works 👇
@MENABCW Day 4 on regulation is the one that matters most right now. The compliance playbook for AI in finance is still being written. Whoever figures out governance without killing innovation speed wins the enterprise market — and probably sets the global standard.
@bestxml007@zothdotio Exactly right. Payment APIs aren't enough. Agents need trust rails — KYC assigned to AI actors, liability frameworks for autonomous decisions, and dispute resolution that works when no human initiated the transaction. The infrastructure problem is harder than the AI problem.
@GBAFReview RAG + compliance is underrated. Most banks still do keyword search for regulatory guidance. Retrieval-augmented generation actually understands context — massive difference for audit trails. Lower cost AND better compliance outcomes. That combination is rare in banking tech.
@Agent_Versity The spending dashboard is just the hook. Once GPT-5.5 starts flagging spending anomalies and suggesting actions grounded in real data, that's when fintech disruption truly begins. The shift from observation to recommendation to action is closer than most banks think.
@charlesmendez Read-only is the trojan horse. 200M users build the habit of asking ChatGPT about their money. Then the 'take action' button appears. Banks aren't losing accounts — they're losing the relationship. The conversation layer is the new distribution moat.
@destraynor@Fin_ai Shopify's payment play is just the beginning. When transaction infrastructure becomes a commodity (API-driven, marginal costs ~0), every digital platform becomes a fintech platform. This is the architectural shift of the decade.
@KeningYe26924 @KobeissiLetter@Fiserv The real shift is happening when AI augments decision-making, not replaces humans. Banking institutions need AI to handle scale, not to reduce headcount. The winners will be the ones who pair AI with better domain expertise.
India is building data centres fast.
65-70% of that capacity is already pre-booked by global hyperscalers.
We're building the pipes. Someone else is controlling the flow.
That's the part we should be talking about.
In 1954, Homi Bhabha told the world that India would build its entire energy future on a metal most countries considered useless.
That metal was thorium. Found in beach sand. India sits on 25% of the global supply.
The problem? You can't just put thorium in a reactor. It's not fissile. It doesn't burn on its own. You have to convert it into uranium 233 first. And to do that, you need a reactor that doesn't exist anywhere outside Russia.
Bhabha knew this. So he designed a 3-stage program where India would BUILD that reactor itself.
Yesterday, India built that reactor.
The man has been dead for 60 years. His plan is still running exactly as written.
Stuck at the Tarmac at the Pune airport in the @IndiGo6E flight since last 4 hours waiting for an empty bay!! Lot of Indigo planes just Parked there and when we suggest solutions they say it’s not allowed !! Unprecedented conditions call for different solutions. @DGCAIndia
The beauty queens joined in with enthusiasm, celebrating #Telangana’s culture by participating in Bathukamma festivities and folk songs, adding colour and charm to the visit.
#IndiaNews
https://t.co/su48O3SbBD
@AirIndiaX we are waiting in the aircraft since last 1 hour for your captains to arrive. What’s happening ?? IX2342. @AAI_Official this is really very disturbing.