Portfolio
Work That Speaks
For Itself.
From AI-powered tools to full-stack platforms and marketing-grade websites — built for real businesses, solving real problems.
Analytics Station
B2B data intelligence platform transforming raw business data into actionable strategic insights for enterprises across hospitality, retail, and healthcare.
Impact
Tech Stack
Overview
Built the complete digital presence for Analytics Station — a data intelligence firm established in 2017, based in Goa. The platform showcases four core service pillars: market research, data analytics, data visualization, and consulting. Designed to convert enterprise decision-makers into clients through credibility signals, measurable outcomes, and detailed case studies. Features interactive dashboards, client testimonials, and industry-specific positioning.
Key Features
Interactive Dashboards
Visual benchmarking tools for competitive intelligence and market positioning.
Case Study Showcase
Detailed project outcomes with measurable ROI across multiple industries.
Client Testimonials
Industry-specific testimonials from operations, hospitality, and healthcare professionals.
Multi-Industry Positioning
Tailored service messaging for hospitality, retail, manufacturing, and healthcare sectors.
ASTECH Corporate Website
Full-stack software agency website with GSAP scroll animations, React Flow workflow diagrams, and a Three.js orbital stack visualization — built to communicate technical depth.
Impact
Tech Stack
Overview
Built ASTECH's own digital storefront — a B2B tech services website covering web/mobile development, AI automation, and infrastructure. Features GSAP-powered horizontal scroll and pin-trigger animations, Lenis smooth scrolling, React Flow interactive workflow diagrams, and a Three.js orbital stack visualization. Designed to communicate technical capability to both technical and non-technical decision-makers simultaneously.
Key Features
GSAP Scroll Animations
Horizontal service scroll, pin triggers, and staggered entry animations throughout.
React Flow Diagrams
Interactive workflow visualizations showing E-Commerce, Sales, and Infrastructure pipelines.
Three.js Orbit Visual
Animated 3D tech stack orbital display in the hero section.
Lenis Smooth Scroll
Custom smooth scrolling with a scroller wrapper for nested GSAP ScrollTrigger support.
ASMR Digital Agency
Research-led digital marketing agency website with 6+ client case studies, 20+ brand logo carousel, and service positioning across 7 growth verticals.
Impact
Tech Stack
Overview
Designed and built the full marketing presence for ASMR — a performance-driven agency that blends data, strategy, and creativity. The site positions the brand around measurable outcomes over vanity metrics. Features dedicated sections for 6+ case studies including Sarovar Hotels, Empress Hill, and Farm Express; an animated client logo carousel with 20+ brands; behind-the-scenes office imagery; and clear service scoping across branding, social media, video production, performance marketing, and advertising.
Key Features
Portfolio & Case Studies
Detailed showcases for Farm Express, Sarovar Hotels, Empress Hill, and more.
Client Logo Carousel
Animated social proof carousel featuring 20+ brand logos.
Interactive Hover Effects
Micro-animations and hover states for a premium agency feel.
Service Architecture
7 clearly scoped service areas each with dedicated positioning and CTAs.
Sugarness E-Commerce
Premium artisanal dessert e-commerce platform with Zomato, Swiggy, and direct ordering integration for a Goa-based confectionery brand.
Impact
Tech Stack
Overview
Built the complete e-commerce and brand presence for Sugarness — a premium dessert retailer handcrafting artisanal sweets in Goa with global influences. The platform integrates Zomato, Swiggy, and direct ordering while telling a brand story built around craft and philosophy. Features a full browsable product catalog, Instagram-integrated social presence, and a detailed "Our Philosophy" narrative that earns trust before asking for a sale.
Key Features
Multi-Platform Ordering
Seamless integration with Zomato, Swiggy, and a direct cart for maximum reach.
Product Catalog
Fully browsable menu with product stories and ingredient transparency.
Brand Storytelling
Philosophy-first design that communicates craft and authenticity before selling.
Social Integration
Instagram feed integration for real-time social proof and product discovery.
Multi-Brand Quotation Platform
One quoting platform serving two manufacturers — domestic catalogue quoting for Exflor, incoterm-priced export quoting for Novacrete — on a single backend that keeps each brand’s data in its own tables.
Impact
Tech Stack
Overview
Two businesses, two very different quotes, one platform. Exflor’s team quotes from a flooring catalogue: pick the customer and site, build the lines with discounts and sq ft / sq m conversion, attach terms, and send a branded PDF that is then tracked from draft through negotiation to an order. Novacrete exports concrete spacers, so their quote is priced from the shipment up — a seven-step wizard where the incoterm decides which freight and local-charge legs apply, packing math converts bags to weight to containers and pallets, and the server returns cost per kilogram, shown to the customer in their own currency while the business tracks it in INR. Both run on one Express and PostgreSQL API with Firebase auth, S3 document storage and per-brand SMTP, with domain tables kept physically separate per brand so a change for one never disturbs the other. A third brand front-end is scaffolded on the same backend.
Key Features
One API, Separate Tables
Shared identity, brand registry and routing; each brand’s products, customers and quotations live in their own tables.
Catalogue Quoting (Exflor)
Product picker with line discounts and sq ft / sq m conversion, and catalogue prices snapshotted so historic quotes never shift.
Export Quoting (Novacrete)
EXW, FOB and CIF each drop the cost legs that become the buyer’s responsibility; palletised and non-palletised packing change the container math.
Server-Owned Pricing
Bags to weight to containers and pallets to freight and cost per kg — calculated on the server, never from client-sent totals.
Branded PDF & Email
Previewed in-app, stored on S3 and emailed through per-brand SMTP, with terms frozen onto the quote as sent.
Versioned & Tracked
Saves write new rows under one reference, and quotes move through a recorded lifecycle that turns into order rows on acceptance.
the problem
Pain points we solved
- 1Quotes were rebuilt by hand each time, so a price could not be reconstructed later.
- 2Export freight and local charges differ by incoterm and were applied inconsistently.
- 3A quote already emailed as a PDF could still be edited in the record.
- 4Two brands needed different quoting logic without running two separate systems.
the solution
What we built
One platform, a quoting flow shaped to each brand.
Separate tables per brand — shared plumbing, independent data.
the process
How it works
- 1Sign in to the brand tool; the API scopes everything to that brand.
- 2Pick the customer, then build the quote — catalogue lines, or shipment, packing and freight.
- 3The server prices it and returns the figures the quote is built on.
- 4Attach terms and review the generated PDF.
- 5Send it, then track versions, negotiation and the resulting order.
AI Location Scouting Tool
AI-driven scouting engine for hospitality expansion — from GPS coordinates to full location intelligence and stakeholder-ready PDF reports in minutes.
Impact
Tech Stack
Overview
Built an AI-powered location scouting engine for Sarovar Hotels' expansion team. Previously, site selection research took days of manual work spread across multiple platforms — no single view of feasibility, demand, transit access, or competition. The tool accepts latitude and longitude coordinates with a scouting radius, then automatically gathers and analyzes data on nearest airports, train stations, and bus depots; top competing hotels; tourism popularity and footfall potential; and area-based hospitality feasibility indicators — synthesizing everything into a structured, decision-ready PDF report for stakeholder alignment.
Key Features
Coordinate-Based Scouting
Enter latitude, longitude, and radius — the engine handles all data gathering automatically.
Transit Hub Analysis
Identifies nearest airports, train stations, and bus depots with distance and accessibility.
Competitor Hotel Mapping
Surfaces top hotels in the area with ratings for competitive benchmarking.
AI Feasibility Report
OpenAI and Gemini synthesize all data into a structured PDF ready for stakeholder review.
the problem
Pain points we solved
- 1Data for site selection is scattered across multiple platforms.
- 2Manual research takes days, sometimes weeks.
- 3No single view of feasibility, demand, and risk exists.
- 4Stakeholder alignment is difficult without clear, structured reports.
the solution
What we built
An AI-driven scouting engine that evaluates locations instantly.
Designed specifically for hospitality feasibility and planning.
the process
How it works
- 1Enter location coordinates.
- 2Define the scouting radius.
- 3AI gathers, analyzes, and benchmarks data.
- 4Generates insights and a structured PDF report.
Centralized Data Pipeline
End-to-end ETL pipeline that autonomously extracts nightly MSSQL backups from 60+ on-premises hotel servers, processes them in parallel via SQS-driven workers on AWS, and surfaces chain-wide analytics on a centralized HQ dashboard.
Impact
Tech Stack
Overview
Sarovar Hotels had zero centralized visibility — each property ran IDS Next PMS on-premises with no connection to HQ. We built a fully autonomous pipeline: PowerShell scripts (multi-version compatible, PS 2.0 to PS 5+) installed via one-time remote access compress the nightly .bak backup to .rar and upload it to a hotel-isolated S3 bucket. A new upload fires an SQS ObjectCreated event that fans out to a horizontally scalable worker fleet on EC2/ECS. Each worker downloads and extracts the .rar, restores the .bak into a staging MSSQL instance, runs an ETL transform, upserts into a centralized PostgreSQL database with one schema per hotel, then ACKs the SQS message and records last_sync_at. The Next.js HQ dashboard reads from PostgreSQL and shows chain-wide occupancy, RevPAR, ADR, booking sources, and sync health per property — with stale-data flags for any hotel that missed a window.
Key Features
Autonomous On-Prem Agent
PowerShell scripts installed once run indefinitely via Windows Task Scheduler — no ongoing remote access or manual steps.
Hotel-Isolated S3 Buckets
Each hotel uploads to its own dedicated S3 bucket. A new .rar upload fires an ObjectCreated event to SQS immediately.
Parallel Worker Fleet
SQS fan-out distributes jobs across EC2/ECS workers. Visibility timeout ensures each message is processed by exactly one worker.
Schema-Per-Hotel Warehouse
Centralized PostgreSQL with isolated schema per property. last_sync_at tracked per hotel — stale properties are flagged on the dashboard.
the problem
Pain points we solved
- 1No hotel was connected to a central system — HQ had zero cross-property visibility.
- 2Hotel servers ran mixed old Windows versions with no guaranteed PowerShell version.
- 3MSSQL .bak files are large and impractical to upload without compression.
- 4Processing backups serially would create multi-hour bottlenecks at chain scale.
the solution
What we built
A fully autonomous, event-driven pipeline from hotel server to HQ dashboard.
Zero manual intervention after the one-time PowerShell installation per hotel.
the process
How it works
- 1Hotel server detects .bak on schedule, compresses to .rar, and uploads to its S3 bucket.
- 2S3 fires ObjectCreated event to SQS — message contains bucket, S3 key, and hotel_id.
- 3Available worker claims the message, downloads the .rar, and extracts the .bak.
- 4Worker runs RESTORE DATABASE into MSSQL staging, then ETL-transforms and upserts into PostgreSQL.
- 5Worker ACKs SQS and writes last_sync_at. HQ dashboard reflects updated chain data.