Prakash Adhikari
Contact

Prakash Adhikari

AI Product Builder

Building AI-powered workflow systems for real desks — LegalTech, document intelligence, and edge-native SaaS.

  • 8 products shipped
  • AI systems built
  • Production software used
Explore my work Start here Contact

Chapter — Proof

Proof

Scale you can feel in daily operations.

What the work has already carried — not vanity metrics.

8
Systems built
6
Live production products
130K+
Trademark records handled
150+
Database migrations
300+
API endpoints
7
Engineering articles
1
Production AI doc workflow
7+
Years experience

Verified from shipped systems only — no vanity user counts. OCR document intelligence is the production AI workflow inside NepalIPMS; more patterns in patterns and the lab.

Chapter — Selected systems

What I build

Selected work with measurable outcomes.

Nepal IP Management System

LegalTech SaaS

Nepal IP Management System

Production IP practice platform for trademarks, patents, deadlines, documents, litigation, and Department of Industry status.

Stage
Production System
Users
Internal Legal Operations
Complexity
High
AI
Integrated
Status
Actively Used
  • 130,000+ trademark records managed
  • 300+ API endpoints
  • 150+ database migrations
  • OCR pipeline for bulletin intake
  • Cloudflare Workers + D1 + R2
  • Used by an IP law firm for daily operations
  • RoleProduct design, architecture, full-stack build
Next.jsTypeScriptCloudflareD1 / R2OCR
Case study nepalipms.com
ScholarQuest

Education · Study abroad

ScholarQuest

Digital platform for destination guidance, test preparation, and counselling workflows.

Stage
Production Platform
Users
Students & Counsellors
Complexity
Medium
AI
Not primary
Status
Live
  • End-to-end student journey from discovery to inquiry
  • Destination + test-prep content architecture
  • Counsellor-ready lead intake
  • RoleProduct design & platform development
Product designWeb platformContent systems
Case study scholarquest.com.np
Hire an Expert Nepal

Marketplace · Talent systems

Hire an Expert Nepal

Marketplace architecture for connecting clients with verified domain experts.

Stage
System Design
Users
Clients & Experts (designed)
Complexity
High
AI
Explored, not primary
Status
Case Study
  • Trust-first verification model
  • Structured brief → match → engagement flow
  • Admin onboarding & quality control paths
  • RoleProduct architecture & system design
MarketplaceMatchingTrust systems
Case study
AI OCR Document Intelligence Engine inside NepalIPMS

Applied AI · Document intelligence

AI OCR Document Intelligence Engine

OCR and extraction pipeline that turns bulletin PDFs into structured matter data — shipped inside NepalIPMS, not as a demo.

Stage
AI Infrastructure
Users
Via NepalIPMS desks
Complexity
High
AI
Core
Status
Prototype → Production
  • PDF → field extraction for legal bulletins
  • Human-in-the-loop confidence review
  • Integrated into NepalIPMS matter intake
  • Edge storage & processing on Cloudflare
  • RoleAI systems design & product integration
OCRWorkersR2Extraction
Case study See in NepalIPMS
TopRank Nepal

EdTech · Public service exams

TopRank Nepal

Bilingual, mobile-first learning platform for Nepal public service exam preparation.

Stage
Production Platform
Users
PSC aspirants
Complexity
Medium
AI
Not primary
Status
Live
  • Nepali / English learning shell
  • Mobile-first study navigation
  • Live production platform for PSC aspirants
  • RoleProduct design & platform build
Mobile-firstBilingual UILearning UX
Case study topranknepal.com
Global Law Associates

Legal · Corporate website

Global Law Associates

Corporate website for IP, corporate law, litigation, foreign investment, and legal knowledge publishing.

Stage
Knowledge Platform
Users
Firm & public readers
Complexity
Medium
AI
Not primary
Status
Live
  • 100+ SEO-ready legal pages
  • Complete CMS for firm publishing
  • Cloudflare deployment
  • Knowledge center for legal content
  • RoleDesign & development of firm digital presence
CMSSEOCloudflareContent architecture
Case study globallaw.com.np
Sri Education

Education consultancy

Sri Education

Educational consultancy platform helping Nepali students with international study opportunities.

Stage
Production Platform
Users
Students & consultants
Complexity
Medium
AI
Not primary
Status
Live
  • Outcome-led consultancy funnel
  • Multi-destination country guides
  • Structured counselling → visa roadmap
  • RoleProduct design & web platform development
Web platformLead intakeContent systems
Case study srieducation.com.np
Yogshala Nepal

Wellness · Brand platform

Yogshala Nepal

Wellness and yoga platform focused on programs, retreats, meditation, and holistic health.

Stage
Brand Platform
Users
Program seekers
Complexity
Low–Medium
AI
Not primary
Status
Live
  • Program & retreat discovery paths
  • Brand-led wellness presence
  • Inquiry-ready responsive experience
  • RoleDesign & development
Brand webProgram UXResponsive design
Case study yogshala.com.np

Chapter — How I think

Featured insights

The strongest thinking — curated.

Not a feed of whatever is newest. The pieces that best show how I research, decide, and ship.

Research → · Engineering → · Principles → · Proof of work →

Chapter — Systems

Product map

How the systems relate.

Hover a product — connected technologies and knowledge light up. Or explore the full body of work.

NepalIPMS

Category
Technologies
Related systems
Open case study

Product explorer

Explore by the kind of system.

Filter by domain — then read problem, users, complexity, architecture, contribution, and outcome.

AI OCR Document Intelligence Engine

Turns messy legal bulletin PDFs into structured matter fields with human review.

See 30-second demo
  • ProblemManual retyping of bulletins into matter records
  • UsersIP attorneys and filing desks
  • ComplexityHigh — noisy PDFs, field confidence, production integration
  • ArchitectureUpload → OCR → extract → review → matter (Workers + R2)
  • ContributionPipeline design, product integration, review UX
  • OutcomeLive intake path inside NepalIPMS

Nepal IP Management System

Full IP practice SaaS — matters, deadlines, register search, litigation, documents.

Case study
  • ProblemSpreadsheets, memory, and fragmented DoIP status
  • UsersIP law firms (daily operations)
  • ComplexityVery high — 130K+ records, 300+ APIs, 150+ migrations
  • ArchitectureNext.js + Cloudflare Workers, D1, R2, FTS search
  • ContributionProduct, architecture, full-stack build
  • OutcomeProduction desk system in active firm use

Global Law Associates

Firm site + CMS + knowledge center for IP and corporate practice.

See demo walkthrough
  • ProblemWeak digital presence for a serious legal practice
  • UsersProspective clients, partners, researchers
  • ComplexityMedium-high — 100+ SEO pages, CMS, publishing
  • ArchitectureContent architecture + Cloudflare deployment
  • ContributionDesign, CMS structure, SEO surface
  • OutcomeLive firm platform at globallaw.com.np

ScholarQuest

Study-abroad consultancy platform — destinations, prep, inquiry.

Case study
  • ProblemFragmented advice for high-stakes education decisions
  • UsersStudents, parents, counsellors
  • ComplexityMedium — content + conversion systems
  • ArchitectureDestination hubs, program modules, structured intake
  • ContributionProduct design & platform development
  • OutcomeLive at scholarquest.com.np

TopRank Nepal / Sri Education

Learning and consultancy platforms for Nepali students.

TopRank case study
  • ProblemExam prep and study-abroad journeys lack structured digital tools
  • UsersPSC aspirants; international applicants
  • ComplexityMedium — bilingual UX, funnel design
  • ArchitectureMobile shells, roadmaps, lead systems
  • ContributionProduct design and build
  • OutcomeLive education products in market

Hire an Expert / UK Community Platform

Marketplace trust systems and an in-progress community platform.

Hire an Expert case study
  • ProblemOpaque expert discovery; community ops without structure
  • UsersClients & experts; community members (in progress)
  • ComplexityHigh — trust, matching, moderation surfaces
  • ArchitectureProfiles, verification, engagement milestones
  • ContributionSystem design; active build on community platform
  • OutcomeArchitecture shipped; community platform currently building

Automation & AI assistants

Deadline engines, intake automation, and AI legal assistant work on the bench.

Currently building
  • ProblemRepetitive legal/ops work that still needs human judgment
  • UsersLegal and operational teams
  • ComplexityHigh — safety, review loops, domain accuracy
  • ArchitectureWorkflow triggers + retrieval + human confirmation
  • ContributionProduct + AI systems design
  • OutcomeProduction OCR/automation; assistants in progress

How I think

AI Product Builder

I design and build intelligent software systems that combine AI, automation, and real-world workflows — not demos that die after the pitch.

AI OCR Search systems Document intelligence Automation SaaS platforms

Thinking essays → · Decisions → · Philosophy →

Working with me

Three clear ways to engage.

Pick the path that matches your problem — then write.

Architecture

Need technical direction? Domain models, Cloudflare stacks, search, OCR pipelines, migration plans.

Ask about architecture →

Consult

Need automation or systems improvement? Constraints, tradeoffs, and a path that fits a small team.

Book a consult →
  • Global Law AssociatesCorporate site/CMS; IP desk also on NepalIPMS
  • Janak Bhandari & AssociatesNepalIPMS production usage (per product register)
  • ScholarQuest · Sri Education · TopRankLive education platforms in Nepal
  • Yogshala NepalLive wellness brand platform
  • TeachingComputer science & web technology instruction

Chapter — Engineering

Product journal

What changed — and what it taught.

Milestones with lessons attached. Full entries on the journal; release list on the changelog.

Aug 2026
v1.0 — AI Product Builder Portfolio

Editorial experience, eight production projects, product map, systems library, writing, knowledge graph, guided journeys — released as a living record.

From here: publish when real work ships — not numbered feature phases. Changelog →

Aug 2026
Reputation layer — insights, proof, principles

Curated featured insights, recommendation paths, founder notes, workflow evidence, contact intents, 2026 report.

Trusted knowledge source — visitors understand how he thinks and works.

Aug 2026
Research, patterns, teardowns, founder brief

Original insights layer: research notes, AI product patterns, design system, search filters, llms.txt.

Recognizable AI product builder brand — ships, thinks, documents, and earns trust.

Aug 2026
Knowledge graph, journal, start paths, diagrams

Related reading across products, decisions, engineering; roadmap, playground, activity.

A continuously growing AI product knowledge platform — not only a complete website.

Aug 2026
Products, search, capabilities, changelog

Product pages, UX in capability matrix, global search, stack, subscribe.

Discoverable AI product builder platform — not only a personal site.

Aug 2026
Lab, Now, Thinking, Profile, product map

Living AI product builder identity: experiments, maturity markers, engagement paths.

Feels active — not an archive of finished work.

Aug 2026
Portfolio v2 — editorial system, case studies, architecture gallery

Evidence-led product builder surface: metrics, diagrams, writing, and live product links.

Public proof of shipping, not a traditional résumé layout.

2026
OCR Document Intelligence Engine in NepalIPMS intake

PDF bulletins → layout OCR → field mapping → human confidence review.

Production AI/OCR workflow inside daily IP firm operations.

2026
Global Law CMS + knowledge center

100+ SEO legal pages, firm publishing, Cloudflare deployment.

Searchable legal content surface for a practicing IP firm.

2025–26
NepalIPMS → Cloudflare Workers, D1, R2

Edge-native rewrite: API surface, document storage, register persistence.

Lower ops burden; infrastructure that matches how the product is used.

2025
130K+ trademark register search for live desks

FTS indexes, normalization, latency budget for phone-side lookups.

Register-scale search usable during client calls.

2025
AI-assisted legal workflow experiments

Retrieval-first drafts with citations; human review before matter files.

Architecture pattern for grounded legal AI features.

2024–25
Education & community product releases

ScholarQuest, TopRank Nepal, Sri Education, Yogshala Nepal shipped live.

Multiple production domains beyond LegalTech.

2024
First production NepalIPMS with active IP firms

Matters, deadlines, documents, and Department of Industry status in one system.

300+ APIs and 150+ migrations under real firm load.

Product timeline

Scroll through the progression.

Foundations → workflow systems → production platforms → AI product systems.

2018

Building foundations

IT associate work — learning how operational systems actually fail under daily load.

2021

Workflow systems

Automation and internal tools that removed repeated operational pain — scripts becoming products.

2024

Production platforms

NepalIPMS, ScholarQuest, education and firm platforms shipping into real desks and markets.

2026

AI product systems

OCR document intelligence, edge-native LegalTech, patterns and research published as a knowledge system.

How I build products

From friction to systems people trust.

Every product starts with a workflow someone is struggling with — then moves through research, craft, and iteration until it disappears into daily use.

Read the full process → · Artifacts →

ResearchSit with the real desk, files, and failure modes
User UnderstandingMap who decides, who operates, and what “done” looks like
Product DesignShape the path users will repeat every day
ArchitectureModel data, boundaries, and failure recovery
DevelopmentAPIs, persistence, auth, and operational safety
AI IntegrationWhere extraction or retrieval genuinely reduces work
DeploymentEdge-native release on Cloudflare
IterationWatch production use; tighten what matters

Technical depth

Tools I ship with — listed, not logo walls.

Architecture · Diagrams · Systems · Full stack page

Cloudflare

  • Workers
  • D1
  • R2
  • KV
  • Pages / Assets

Application

  • Next.js
  • React
  • TypeScript
  • SQLite
  • Postgres

Intelligence

  • AI APIs
  • OCR Systems
  • Vector Search
  • Workers AI

Craft

  • Git
  • GitHub
  • Docker

Chapter — Current focus

Current work

Active work on the bench.

Now page · Roadmap · Lab

  • AI-native workflows
  • Document intelligence
  • Legal technology
  • Product infrastructure

Living Now page → · The Lab →

Available for

Focused collaborations.

  • AI Product Development
  • Technical Architecture
  • SaaS Consulting
  • Automation Systems
  • Product Strategy

I build useful software — systems that become part of someone’s daily workflow.

Engineering, design, and business understanding in one practice. The best products disappear into the background; users remember the outcome, not the interface. Read the philosophy

Portrait of Prakash Adhikari

About

From IT systems to applied AI.

I build and ship production-grade digital products — SaaS platforms, automation systems, and domain-specific tools — with a particular focus on legal technology and education.

Today that practice extends into applied AI: retrieval, document intelligence, and agents that amplify expert judgment rather than replace it.

Product Designer Software Architect AI Engineer

Teaching / Experience

Sharing the craft.

Teaching sharpened how I explain systems — and how I design products people can actually operate.

  • Computer Science teachingClassroom instruction in foundational CS and problem-solving.
  • Web technology mentoringPractical web engineering shared with students and peers.
  • Product knowledge sharingArchitecture, shipping, and domain-shaped product thinking.
  • Technical educationApplied sessions on intelligent systems and how products get built.

Portfolio assistant

A guided explorer — not a chatbot.

Answers use only published portfolio pages. No external AI. No invented claims. Privacy-friendly.

Ask me about

A human touch before you write.

Topics that usually lead to useful conversations — then contact with context.

  • AI product developmentFrom desk friction to a shippable, reviewable slice.
  • Legal technologyMatters, deadlines, registers — software shaped like IP desks.
  • OCR systemsDocument intelligence where the review screen is the product.
  • Cloudflare architectureWorkers, D1, R2 under production constraints.
  • Building SaaS productsDomain models, migrations, and systems small teams can run.

Go deeper

Architecture, writing, gallery, open source.

Start here Assistant Ask me about Featured insights Compare Status Evidence Tech evolution For founders Founder notes Proof of work Knowledge library AI Product Playbook Resources Lessons How I build Artifacts Principles Workflows Contact 2026 report Products Research Patterns Teardowns Design system Journal Roadmap Diagrams Playground Contributions Activity Changelog Capabilities Stack Now The Lab Thinking Profile Subscribe Decisions Architecture Engineering Gallery About

Chapter — Let’s build

Contact

Let’s build something intelligent.

Based in Kathmandu. Choose an intent — build, improve, architecture, or collaborate — then write with context.

  • AI Product Development
  • Technical Architecture
  • SaaS Consulting
  • Automation Systems
  • Product Strategy

Open contact →

hello@prakashadhikari.dev