~/portfolio
about.md
work/
personal/
experience.log
contact.mk
UTF-8 · LF · v2026.04
open to work
dulguun~ $whoami --verbose

dulguun.
ship(boring)
// until it isn't

Developer at Sendly NBFI, Mongolia. I build Odoo ERP and web systems that automate loan workflows, using React, Next.js, Python, and JavaScript.

→ cat work/→ git log↓ resume.pdf$ mail me
##

01 / about.md

287 words · 2 min read

I started out as an electrical engineering student at the National University of Mongolia, then moved to Tokyo for two years at Yu Language Academy — long enough to pass JLPT N2.

Back in Ulaanbaatar in 2023 I joined Sendly NBFI on the IT support team. I spent nights teaching myself React and shipped the company's first employee directory. That got me onto the software team in August 2024, and I've been building the core loan system since.

I like quiet software — the kind that sits inside a company, runs a process that used to eat three people's afternoons, and doesn't need marketing to exist. Most of what I've built is like that.

./dulguun.tomlREAD-ONLY
nameDulguun PurevtserenroleSoftware Developercompany@ Sendly NBFIsince2023-05 · 3 yrseduNUM · BA Elec. Eng.tokyoYu Lang Academy · N2emaildulguunnu@gmail.comwebdulgx.com
##

02 / work/ featured

8 systems · shipped
● LIVE · DAILY USEBEFOREExcel.xlsx3 sheets · 40+ formulasmanual re-keymigrate()AFTERLoanCore.pyscoring · DTI · eligibilityautomated · audited · loggedINTEGRATIONSHURgov. data exchangeKYCauto-verify customerCollateral3rd-party valuationE-Barimttax-receipt pipelineOUTPUTSdecisions · committee votes · tax receiptsgenerated, logged, archived — no human re-keyingIMPACTmanual labor on loan processing: −50 to −60%loan approval cycle time: days → hourscustomer data entry errors: near zerocommittee decisions: 100% digital · auditable
FEATURED · 2024—present · ● sendly-nbfi

A loan system that doesn't need
Excel anymore.

I rebuilt Sendly's core loan pipeline inside their ERP: application intake, credit scoring, eligibility math, and debt-to-income ratios — everything that used to live in three spreadsheets across four desks.

Then I wired it outward: HUR for government data, KYC for verification, a digital loan committee with voting and minutes, collateral management with third-party valuation, and E-Barimt for tax receipts.

−50–60%
manual labor
6
systems · one ERP
daily
active users
dulguun~/work $ls -la --sort=impact
#datetypeprojectstack
01
2024 · Q1–Q2
system
Loan Decision Automation
End-to-end automation of loan application to decision flow — intake, scoring, risk checks, and approval routing, replacing manual Excel-based workflows.
OdooPythonPostgreSQL
02
2024 · Q2–Q3
integration
Credit Scoring (Sain Score · Buren Score)
Integrated external credit scoring systems into the loan pipeline — automated fetching, evaluation, and decision support using Sain Score and Buren Score.
OdooPythonREST
03
2024 · Q3
web.app
Sendly Intranet
Employee directory, branch contacts, news feed, Knowledge Base, suggestions — centralized everything that used to live in email and Excel.
Next.jsReactTailwind
04
2024 · Q4
integration
HUR · KYC pipeline
Plugged the loan system into Mongolia's government data-exchange backbone to auto-fetch and verify customer information.
OdooPythonRESTFastAPIPostgreSQL
05
2025 · Q1
module
Loan Committee
Digitized committee decision-making — member voting, meeting minutes, auditable decision approval, full search.
OdooPostgreSQLXML
06
2025 · Q2
module
Collateral Management
Registration and valuation module with third-party integrations — improved data accuracy and valuation speed.
OdooPythonAPIsAWS S3
07
2025 · Q4
integration
E-Barimt Tax Pipeline
Connected the loan pipeline to Mongolia's tax receipt system — every transaction auto-generates and submits its own И-Баримт.
OdooPythonTax APIAWS LambdaAWS SQSPostgreSQL
08
2026 · Q1
web.app
Loan Mini App (Frontend Contribution)
Contributed to the customer-facing loan mini app — improved UX flows, application steps, and data validation for smoother loan requests.
Next.jsReactTailwind
##

03 / personal/ featured

1 project · in use
● LIVE · DAILY USEMODULESSRS Flashcardsspaced repetitionSmart Quizadaptive · multi-choiceKanji Tracerstroke-order · canvasFEATURESMatching Gamestimed · active recallAI Story GenGemini · vocab in contextAcademic Newscurated · annotatedDATASupabase · Auth + DB + realtime statsprogress tracking · review history · SRS intervalsGOALmaintain fluency → achieve JLPT N1
PERSONAL · app.dulgx.com ↗

A Japanese learning app built for myself
to reach N1.

Mobile-first web app with spaced repetition flashcards, adaptive quizzes on kanji readings and usage, and hands-on stroke-order tracing on canvas.

Timed matching games for active recall, an AI story generator (Gemini) that drops target vocab into reading passages, and a curated academic news view — all backed by Supabase with full progress tracking.

SRS
smart scheduling
6
learning modes
N1
target level
dulguun~/personal $ls -la --sort=impact
#datetypeprojectstack
01
2025 · Q2
web.app
Japanese Learning App ↗
Mobile-first SRS flashcards, kanji tracer, adaptive quizzes, matching games, AI story generator, and academic news — built to reach JLPT N1.
Next.jsTypeScriptSupabaseGemini
##

03 / git log --experience

2 commits
commit 1f8a2c2024-08-01 → HEADHEAD

feat: move from support to software team

Software Developer @ Sendly NBFI · Ulaanbaatar
  • Dev on the core loan automation system, replacing Excel-based workflows with an ERP pipeline — 50–60% reduction in manual labor.
  • Shipped integrations with HUR (gov. data), KYC, E-Barimt (tax), and third-party collateral valuation.
  • Built the loan committee module (voting, minutes, approval) and the collateral management system (registration, valuation).
  • Shipped the intranet portal end-to-end in Next.js — directory, news, Knowledge Base, suggestions.
commit 4b2d012023-05-10 → 2024-07-01

init: first job after Tokyo

IT / Support Engineer @ Sendly NBFI · Ulaanbaatar
  • Administered Active Directory, Office 365, email, financial software, and internal web apps at permission level.
  • Remote and on-site technical support to UB HQ and countryside branches.
  • Maintained network stability and upheld information-security policy.
  • On my own initiative, taught myself React and shipped Sendly's first employee directory — the project that got me onto the software team.
##

04 / package.json

20 deps · 0 bloat
{"name": "dulguun","role": "full-stack · erp · frontend","since": "2023",
"backend": {"odoo": "^16.0", // main ERP"python": "3.x","postgresql": "*","xml-rpc": "*"},
"frontend": {"next": "^15.0","react": "^19.0","tailwindcss": "^4.0","javascript": "*"},
"infra": ["git", "gitlab", "azure-devops", "aws"],"ops": ["active-directory", "office-365"],
"languages": {"mn": "native","en": "professional","ja": "JLPT N2"}}
##

05 / off-the-clock

[ 35mm ]
01 · Photography

Mostly film and live concerts. Good practice for noticing what's worth waiting for.

[ EE ]
02 · Electronics

My degree was in electrical engineering; I still like pulling things apart to see how they work.

[ ♪ ]
03 · Live music

If a band is in UB I'm probably at it. Tokyo spoiled me for small venues.

[ N2 ]
04 · Languages

N2 earned the slow way — by living there. English came from the internet. Mongolian came for free.

dulguun~ $mail --start-conversation

$ dulguunnu@gmail.com
— or just wave.

EMAIL
dulguunnu@gmail.com
SITE
dulgx.com
STATUS
● open to work