Billeasy · 2026Transit & payments
MMB Ferry Ticketing
Digital ferry ticketing for Maharashtra Maritime Board on a multi-tenant transit platform.
Next.jsNode.jsTypeScriptMongoDB
An open book
Akshat Jainin eight chapters.
Senior Software Engineer with 4+ years building full-stack, data-heavy and AI-native products. Currently shipping transit ticketing at Billeasy for about 150K daily transactions. Everything below is interactive; each chapter links back to the chat if you want to dig deeper.
Chapter 01
A day in my life
Here's how a typical 9-hour day splits for me as an AI Native Software Engineer at Billeasy. Toggle to see how it shifts on a release day.
Donut chart of a 9-hour workday. Regular day: building features 3.5h, reviewing diffs 1.5h, system design 1h, debugging 1h, learning 0.75h, syncs 0.75h, deploys 0.5h. Release day: debugging 2.5h, building 2h, reviewing 1.5h, deploys 1.5h, syncs 1h, design 0.5h.
9ha regular day
On regular days most of my time goes into building and reviewing. On release days debugging, verification and deploys take over, because a ticketing platform with real payments has to ship safely.
Chapter 02
How I rate my skills
This is my honest self-assessment. 10 means I could teach it, 7 means production-confident, 5 means working knowledge. Pick a category to see the individual skills.
Radar chart of six skill categories rated out of 10: Frontend 9, Backend 8, Databases 8, System design 8, AI and LLM 7, DevOps and Cloud 6.
Self-assessed · 10 = could teach it · 7 = production-confident · 5 = working knowledge
Frontendoverall 9/10
React9/10
Data viz (D3, Plotly, Recharts, Kepler)9/10
Next.js8/10
TypeScript8/10
Tailwind / shadcn8/10
State (Redux, React Query, Zustand)8/10
My strongest areas are frontend and data-heavy UIs, backend API design and database performance. DevOps and classical ML are where I'm still building depth.
Chapter 03
What I value, and what I am working on
These are the values I actually work by, each with a real example. Next to them: my strengths, and the things I am honestly still working on.
Six value cards: ownership, depth over hype, users over specs, integrity, AI as a multiplier, good energy in teams. Strengths listed next to growth areas with what I am doing about each.
Strengths
- System design thinking
- Performance optimisation
- End-to-end ownership
- Working under pressure
- Written communication
What I'm working on
Perfectionism slowing execution
→ Time-boxing design, then shipping behind flags and iterating.
Verbal communication
→ Writing things down first, then presenting them; speaking up earlier in design reviews.
Deep ML and LLM internals
→ Studying how transformers and embeddings work, not just the APIs.
Large-scale data systems
→ Going deeper on streaming, partitioning and query engines like Trino.
Ownership and depth over hype drive most of my decisions. I'd rather tell you what I'm still improving than pretend I have no gaps.
Chapter 04
How I plan and execute
This is the loop I follow for any feature, from a small fix to a new product like the MMB ferry ticketing. Pick a step to see what I do, which tools I use and a real example.
Seven-step process: Understand 10%, Design 15%, Stress-test 10%, Build 30%, Validate 15%, Ship 5%, Observe 15% of effort.
Share of effort per step
Understand~10% of effort
Align with product, ops and UI on the real problem, users and constraints before any code.
NotionFigmaConversations with users
Real exampleSat with the Customer Success team to learn how they actually analysed campaigns every day.
I spend a lot of effort before and after the code: understanding, designing and stress-testing up front, then validating and observing in production. AI speeds up the build step, but the thinking stays mine.
Chapter 05
Experience and domains
My career from 2022 to today. Switch the lens between roles and domains: EdTech, travel and fintech services, marketing analytics, and transit ticketing and payments.
Timeline: SI Online intern Jan to Apr 2022 (travel and fintech services), Finemate founder May 2022 to Jul 2023 (EdTech), Infinite Analytics software engineer Aug 2023 to Jul 2026 (marketing analytics and BI), Billeasy AI Native Software Engineer Jul 2026 to present (transit ticketing and payments).
20222023202420252026
Travel & fintech servicesEdTechMarketing analytics & BITransit ticketing & payments
AI Native Software Engineer at Billeasy
Jul 2026 – Present · Transit ticketing & payments
Multi-tenant metro, bus and ferry ticketing: 25M+ PWA hits, ~150K daily transactions; owned MMB ferry end-to-end.
Four domains, one common thread: I own products end-to-end and make them fast and reliable at real scale.
Chapter 06
What I know in AI
My AI knowledge, split honestly into three tiers: what I've shipped in production, what I've built in projects, and what I'm still learning.
AI skills in three tiers. Shipped in production: multi-agent NL-to-SQL system, RAG, AI-native development with Claude Code and MCP, prompt engineering. Built in projects: LLM-as-judge grading, local open-weight LLMs, classical ML spend recommendation. Learning: Pinecone, Neo4j, Airflow, LlamaIndex, LLM internals.
Shipped in production6
Used by real users or a real team
Built in projects3
Working projects, not yet at production scale
Learning6
Actively studying; no expertise claimed
I've shipped RAG and multi-agent systems and I work AI-native every day. I only move something into 'shipped' once I've run it for real.
Chapter 07
What I'm looking for
I'm looking for a Senior Software Engineer role, hybrid or remote, in Mumbai, Bangalore, Pune, Indore, Ahmedabad, Jaipur, Noida or Gurgaon. I'm open to any company stage. Here's what matters most to me.
Priorities out of 10: ownership and impact 10, hard technical problems 9, growth 9, respectful culture 8, sustainable pace 8, AI and data products that help people 7. Senior Software Engineer, hybrid or remote, eight Indian cities, any company stage.
- Role
- Senior Software Engineer · Hybrid or remote
- Locations
- 8 cities in India
- Company stage
- Open to any company stage
MumbaiBangalorePuneIndoreAhmedabadJaipurNoidaGurgaon
What matters most · weight out of 10
Ownership & impact10/10
I do my best work when a system is mine end-to-end.
Hard technical problems9/10
Scale, AI and data problems keep me sharp and energised.
Growth & learning9/10
I want to keep levelling up, technically and as a leader.
Respectful, low-politics culture8/10
Sharp people, honest feedback, shared ownership.
Sustainable pace8/10
Intense sprints are fine; permanent firefighting is not.
Products that help people7/10
Like transit ticketing: software millions rely on every day.
Must-haves
- Real ownership of features and systems
- A respectful, low-politics culture
- A sustainable pace with real work-life balance
- Room to grow into senior technical leadership
Nice-to-haves
- AI-native engineering practices
- Data-heavy or AI/ML products
- Products with real-world impact at scale
- A fast-moving team that ships often
Ownership, hard problems and a healthy culture matter most. I'm fine with intense sprints; permanent firefighting and a toxic culture are the only things I rule out.
Chapter 08
Projects and impact
Here are the measurable wins first, then every project I am proud of. Filter by domain or tech, and open a card for my role and key decisions.
Before and after chart: geospatial report load 300s to 15s, campaign dashboard load 180s to 30s. Ten projects: MMB ferry ticketing, Metro PWA and payments, Bus commuter PWA for MSRTC, Purple and Neeta, Bus operator dashboard, Sherlock geospatial dashboard, Sherlock monitoring dashboard, Khuli Kitab, Unhyped (an interactive, fact-checked history of AI), Finemate 2.0 (a social-first dating app matching on communication style), and Finemate (a spaced-repetition learning platform).
Measured impact · load time in seconds
BeforeAfterBilleasy · 2026Transit & payments
Metro Commuter PWA & Payments
One codebase serving Mumbai, Hyderabad, Nagpur, Pune and Noida Metro commuters.
25M+ PWA hits · ~150K txns/day
Next.jsReactRedux ToolkitNode.jsBilleasy · 2026Transit & payments
Bus Commuter PWA
White-label bus ticketing PWA for MSRTC, Purple and Neeta: booking, seats, passes, QR tickets.
Part of 25M+ PWA hits · ~150K txns/day
Next.jsReactTanStack QueryFastifyBilleasy · 2026Transit & payments
Bus Operator Dashboard
Operator dashboard for MSRTC and Neeta: routes, fares, passes, dispatch and operators.
Next.jsReact 19Tailwindshadcn/uiSherlock AI · 2024–26Analytics & BI
Geospatial & Audience Dashboard
Explore GB-scale location and visitation data on maps and build niche audiences by joining datasets.
5 min → 15 s
ReactKepler.gldeck.glFastAPISherlock AI · 2023–25Analytics & BI
Monitoring & Analytics Dashboard
A Power BI-like platform comparing 30+ KPIs across Google and Meta campaigns.
3 min → 30 s
ReactFastAPIMySQLRedisPersonal · 2025–26AI
Khuli Kitab
This chatbot: a RAG knowledge base over my documents, with charts like this one.
FastAPILangChainChromaDBGeminiPersonal · 2026AI
Unhyped
The history of AI as a road you travel backwards in time, with every hyped claim checked against reality.
354 sourced entries · 57 claims fact-checked
Next.jsReact Three FiberTypeScriptZodFounder · 2026Consumer social
Finemate 2.0
Social-first dating app for India that matches people on how they text, not how they look.
~62K lines of TypeScript, built solo
React NativeExpoHonoPostgreSQLFounder · 2022–23EdTech
Finemate
A "retention engine": social learning built on spaced repetition and gamification.
ReactFlaskMongoDBMaterial UI
Across projects the pattern repeats: slow, manual workflows turned into fast, self-serve products. The geospatial reports went from 5 minutes to 15 seconds.