Open to new conversations
Akshat Lambhate
Founder's Office · Business Automation · LiteStore
I build the systems a retail business runs on.
Pipelines that ingest every store's sales with nobody in the loop, the dashboard leadership makes renewal calls from, and an AI assistant that reads the lease agreements. Each one replaced a vendor or a spreadsheet. Each one is in daily use.
- of sales aggregation automated
- ~98%
- brands reported, unattended
- 100+
- reported downtime
- 0
- Next.js
- React
- Node
- Prisma
- TypeScript
- Three.js
- Tailwind
- PostgreSQL
- Supabase (pgvector)
- AWS Bedrock
- RAG & multi-agent
- n8n
- AWS EC2
- Docker
- Dokploy/Traefik
- Cloudflare Tunnel
- GitHub Actions CI/CD
Experience
Four roles, one instinct: find the manual step and retire it.
Each position ended with something that used to be done by hand running unattended. The latest one owns the entire technology layer of a retail business.
- Sept 2025 — PresentCurrent
Business Automation Lead, Founder's Office
LiteStore·Bangalore, India

Own the technology layer end to end: the data that arrives, the platform leadership decides from, and the AI that reads the contracts.
- Architected an orchestrated sales pipeline (n8n, SFTP, POS APIs → PostgreSQL) that ingests every store’s data with no human input, reaching ~98% automation of monthly aggregated sales as a single auditable source of truth.
- Deployed an unattended daily/weekly/monthly reporting engine serving 100+ partner brands across the FLXY and THE RACK formats, absorbing the operations team’s full reporting workload with no downtime.
- Built a central operations dashboard (Next.js, Prisma, PostgreSQL) spanning store sales, lease commercials and MBO performance, with auto-computed PSFPD and a store-level Gross P&L used by Finance and leadership for renewal and exit decisions.
- sales aggregation automated
- ~98%
- brands auto-reported
- 100+
- paid tools replaced
- 3
- n8n
- Next.js
- Prisma
- PostgreSQL
- AWS Bedrock
- Supabase
- Zoho Books
- June 2025 — Aug 2025
Software Developer Intern
LiteStore·Bangalore, India

Laid the data foundation the rest of the platform was later built on.
- Built the core data structure and a live store-portfolio dashboard, automating sync to Google Sheets and cutting data latency by 90% via API-driven updates.
- lower data latency
- 90%
- JavaScript
- Google Sheets API
- Apps Script
- Apr 2025 — June 2025
Software Developer Intern (Automation)
Incentiv·Remote

Automated a content operation end to end, from creative to distribution.
- Built end-to-end social media automation in n8n, Make and Lindy with AI-generated creatives, plus auto-scheduling and cross-platform distribution, cutting time-to-publish by ~70%.
- faster time-to-publish
- ~70%
- n8n
- Make
- Lindy
- Generative AI
- RSJan 2024 — May 2024
Front-End Developer Intern
RSL Soft Tech Pvt. Ltd·Indore, India
First production codebase. Where the MERN fundamentals came from.
- Gained hands-on experience in the MERN stack; implemented secure JWT authentication and authorization mechanisms for client projects.
- React
- Node.js
- Express
- MongoDB
- JWT
Selected work
Six systems, all of them in daily use.
Internal platforms, data pipelines, an AI assistant and two brand sites. Real screens throughout — nothing here is a mockup.
Operations & Revenue Dashboard
A single cockpit for store performance and lease commercials
The central operating picture for the business — every store’s sales, the commercials of every mall and outlet lease, and MBO performance in one place, updating from the automated data layer beneath it.


- Built
- A Next.js + Prisma + PostgreSQL dashboard spanning store-level sales, mall/outlet lease commercials and MBO performance, with auto-computed PSFPD, KPIs and trend analysis.
- Extends into
- LS-PMS — a Project Management System layered onto the same dashboard for the Projects arm, across core, workflow, automation and portal modules with shared authentication.
- Used for
- Renewal, exit and renegotiation decisions taken directly off the store-level views by Finance and leadership.
- source of truth for ops
- 1
- auto-computed per store
- KPIs
- stores under one platform
- 100+
- Next.js 15
- Prisma 6
- PostgreSQL
- Custom JWT auth
- Recharts
Automatic P&L Engine
Store-level profit & loss, computed without touching a sheet
Profitability per store used to live in manual spreadsheets assembled after month-end. This engine computes a store-level Gross P&L automatically from the live sales data and lease commercials already flowing through the platform.


- Built
- An automatic Gross P&L view joining each store’s sales against its month-wise lease commercials and cost lines, recomputing as new POS data lands — no manual consolidation step.
- Feeds
- The same renewal / exit / renegotiation decisions leadership runs off the dashboard, now with profitability attached, not just top-line sales.
- Backed by
- The lease module — a single system of record holding terms, month-wise commercials and executed agreements for every store, replacing an externally procured tool.
- manual consolidation steps
- 0
- recompute on new POS data
- Nightly
- PostgreSQL
- Prisma
- Automated joins on live POS data
- Lease commercials module
LEXA — RAG Legal Assistant
A multi-agent chatbot that reads lease agreements and answers in plain language
Retail lease agreements and revenue-share agreements are dense and repetitive to query by hand. LEXA retrieves the right clauses and answers questions against them, built and hardened across multiple versions.

- Architecture
- A supervisor-plus-tools multi-agent design on AWS Bedrock (Claude Sonnet + Haiku), with Supabase pgvector for retrieval — built in n8n and iterated v4 → v5.
- v5 added
- Postgres-backed chat memory and scoped vector retrieval, so answers stay grounded in the correct document set rather than the whole corpus.
- Alongside it
- Two organisation-wide Generative AI workshops for the Design and Marketing teams — now part of their day-to-day workflow.
- iterations hardened in production
- v5
- org-wide AI workshops run
- 2
- AWS Bedrock (Sonnet + Haiku)
- Supabase pgvector
- n8n
- Multi-agent (supervisor + tools)
- Postgres chat memory
Sales Data Pipelines & Automations
The automated backbone that reports revenue with no human input
Every downstream system — the dashboard, the P&L, the revenue-share numbers — depends on clean sales data arriving reliably from dozens of stores and POS systems. This is the orchestration layer that makes that happen automatically.

- Ingestion
- n8n workflows pulling from multiple POS systems (Wondersoft, eShopaid, SOAP APIs) and SFTP brand feeds, consolidating multi-store data from brands like Kisah, Bewakoof and FLXY into Supabase.
- Logic
- Automatic application of Net Sales and Retail Sales definitions plus GST consolidation — reaching ~98% automation of monthly aggregated sales as a single auditable source of truth.
- Reporting (ADSR)
- An unattended daily / weekly / monthly reporting engine serving 100+ partner brands across FLXY and THE RACK, with a validation layer flagging duplicate transactions and missing sales dates.
- Messaging
- A WhatsApp automation pipeline (Evolution API on EC2 + n8n + Chatwoot) for omnichannel CRM, and a WhatsApp-to-daily-report flow for the Projects team.
- aggregation automated
- ~98%
- brands auto-reported
- 100+
- reported downtime
- 0
- n8n
- Supabase
- SFTP
- POS APIs
- Evolution API
- Chatwoot
- Apps Script
STRETCH Premium Outlets — Website
A scroll-driven brand site for LiteStore’s premium-outlet format
STRETCH is LiteStore’s premium outlet-centre format. The site was ideated and built fully in-house to an Awwwards-grade standard — a scroll-driven experience carrying the brand’s signature “stretched letterform” device.

- Craft
- A complex SVG clipPath wordmark effect, a day/night theme system, and scroll-linked motion — built on Next.js and being extended into Framer sections.
- Brand system
- Space Grotesk + Lato type; Stretch Gold (#FDB52E), Midnight Navy (#101342), Retail Red (#D13220); tagline “Great Brands. Great Prices. / Worth the drive.” First location: Attibele, Bengaluru.
- AI production
- Veo 3 brand-specific prompt packs for AI video, a Claude Code skill package for the build, and an AI creatives production guide.
- in-house, no agency
- 100%
- Next.js
- SVG clipPath / motion
- Framer
- Day-night theming
- Veo 3
LiteStore — Website
Audited, rebuilt and owned in-house — no agency, no retainer
The primary LiteStore web property — audited, taken in-house and now maintained directly, removing external agency development and retainer cost. Managed alongside the FLXY site.

- Ownership
- Full control of the codebase and content, with updates shipped in-house instead of routed through an agency cycle.
- Sibling property
- The FLXY brand site — audited and managed under the same in-house model.
- brand sites taken in-house
- 2
- Next.js / React
- In-house CI/CD
- Cloudflare
- Self-hosted infra
Education & credentials
An engineer's training, pointed at operating problems.
A computer science degree, an internship that turned into a Founder's Office seat, and the credentials collected on the way.
- 2018Higher SecondaryShri Vaishnav Academy78.6% · 83%
- 2020B.Tech, Computer ScienceSVVV, IndoreCGPA 8.09
- 2024Front-End Developer InternRSL Soft TechFirst production codebase
- 2025Automation InternIncentivA content operation, automated
- 2025Founder's OfficeLiteStore, BangaloreOwns the technology layer
- 2020 — 2024Indore, India
Shri Vaishnav Institute of Information TechnologySVVV
B.Tech, Computer Science Engineering

CGPA 8.09
- 2018 — 2020Indore, India
Shri Vaishnav Academy
Higher Secondary & High Secondary

78.6% · 83%
Certifications
1 held
Founder's Office Fellowship
Outside the work
Seven things I give real time to.
Most of them are the day job in disguise — take something apart, learn why it works, put it back together better.
01 / 07Playing Guitar
Indie folk, mostly. The fastest way I know to stop thinking about work.
The artifact
The one-page version.
Everything above, on a single page. The file starts downloading the moment you click — the animation is only the receipt.
- 1 page
- 125 KB
- Updated September 2026

