How We Hire at SoduraAI: Building a Team for Mongolia's Legal Future
We are not building a standard software company. We are building Mongolia's legal AI infrastructure — and that changes who we hire, where we look, and how we measure people.
Mongolia needs its own legal AI.
Not a translated one. Not a version of a Silicon Valley product slightly adapted for local language. An AI built on Mongolian law, Mongolian court rulings, Mongolian legal culture, and Mongolian users' habits.
That is what Huuli means when we say the product has to be local-first.
The same truth applies to the team.
When we hire, we are not just filling seats. We are recruiting people who can build something that earns the legal profession's trust, that respects the gravity of law, and that works in the unique conditions of Mongolia's legal market.
Here is how we think about hiring at SoduraAI.
The hiring context is unlike anywhere else
Mongolia is a small market with outsized ambition.
The startup ecosystem here is growing — rank #86 globally in 2025, plus 75 startups, with ecosystem growth over 30% in a single year. That is real momentum. But the talent pool is small, and the market is not large enough to absorb that talent until companies create the roles.
For an AI company working on legal domain expertise, the challenge is sharper. We need people who can write production AI systems and people who can read, understand, and reason about Mongolian law.
That combination is rare.
In most countries, you would split legal and technology cleanly: lawyers know law, engineers know code. In Mongolia at this stage, the percentage of engineers who also have legal interest — or lawyers who are technically curious — is even smaller than in mature markets.
We do not have the luxury of hiring for a single specialty and hoping the rest fills in. We need a team that can move fast across the boundary between technology and legal reasoning.
The roles we actually need
When you build a legal AI, the team is not just "engineers and sales people."
1. AI / NLP Engineers The core problem is Mongolian language understanding layered on a large, structured legal corpus. We need people who can work with RAG, semantic search, document processing, citation accuracy, and fine-tuning — specifically for Mongolian and Japanese. This is not generic NLP work; it requires comfort with low-resource language models and the ability to build reliable retrieval pipelines.
2. Legal Domain Context We need people who can translate court rulings and statutes into evaluation criteria. This can be a practicing lawyer, a legal researcher, or a law librarian — but in every case, the role is to ensure accuracy before errors reach users. Legal AI is different from a chatbot: when a lawyer relies on your answer in court preparation, the margin for hallucination is zero.
3. Product / Service Engineers The jobs page calls this "professional AI services for liability-intolerant environments." That phrase is deliberate. We sell to the Mongolian Supreme Court, to law firms, to compliance-heavy clients. Someone on the team needs to understand deployment, trust architecture, and how to demonstrate accuracy to technical buyers who are naturally skeptical.
4. Cross-Market Engineers Because we run in both Mongolia and Japan, we need engineers comfortable with bilingual or multilingual product thinking. Some of our team are educated in Tokyo; some have lived in Japan. That cross-cultural fluency is itself a competitive advantage.
5. Growth / Business Development The legal AI buyer in Mongolia is not the same as a SaaS buyer in San Francisco. The sales cycle is longer, the buyers are professionals with strong referral networks, and the product must earn trust in person before a contract is signed.
Where we look
In Mongolia: Mongolia's engineering talent is concentrated at universities like the National University of Mongolia and MUST, and at local firms like Bolorsoft. But the best candidates are often not on LinkedIn first; they are active in Mongolian tech communities, on Facebook, and through personal networks. The cadence of hiring here still leans heavily on community and reputation.
In the diaspora: Some of Mongolia's strongest engineers are in Tokyo, Europe, and the U.S. Many Mongolians abroad want to participate in building the country's infrastructure without relocating. We actively engage the diaspora through remote-friendly roles — particularly engineers who already have RAG/AI experience and technical fluency in Mongolian.
In Japan: Because we have products like Roppolab and partnerships like Houkaitei (Japanese pharmacists), a small Japan-facing team matters. Japan-side hiring is not our primary market, but bilingual engineers and Japan-experience engineers are real leverage.
What we screen for beyond résumé:
- Curiosity about law. Even if a candidate is not a lawyer, they should ask questions about how legal research actually works.
- Grit with ambiguity. We build infrastructure where both the technology stack and the market conditions are under construction.
- Comfort with accuracy culture. Law does not reward speed over precision.
- Pride in Mongolian context. We want people who treat Mongolian legal AI as a first-class engineering problem, not a side project.
What makes hiring here different from a Singapore or San Francisco startup
1. The market is small, so mission matters more In a large market, people can optimize for salary and lifestyle. In Mongolia, early-stage company hires face higher risk but also unusually high impact. We lead with purpose: building something that did not exist before.
2. Brain drain is real — but it cuts both ways Talented Mongolians are more likely to be in Tokyo than in Ulaanbaatar. That is a retention problem for many companies. For us, it is partly a distribution advantage: we have both Mongolia and Japan as natural markets, and Mongolian engineers abroad have an immediate bridge to both.
3. Trust is earned step by step Because our customers are courts, lawyers, and compliance-heavy organizations, credibility matters more than hype. We hire people who communicate carefully, who treat accuracy non-negotiably, and who are comfortable with slower, relationship-driven sales cycles.
4. Startup salaries are not the same as global benchmarks Mongolian engineering salaries currently run from around $5,000 to $11,000 annually for entry-level roles. That is significantly below what the same skill set commands abroad or even in other regional markets. We backfill some of that gap with equity, ownership, and mission alignment — but we are also honest: not every stage of the company can match global Silicon Valley compensation. We hire people for whom that trade-off is worth it.
The early team structure we are moving toward
The founding team already covers AI engineering and business development across Mongolia and Japan. As we grow, the first added roles tend to be:
- Legal AI Research Engineer: Someone who can build eval frameworks, work with legal documents, and feed the product loop with real Mongolian legal data.
- Legal Domain Specialist: To validate outputs, maintain court data pipelines, and shape the product from a practicing-legal-experience angle.
- Backend / Platform Engineer: To harden API infrastructure and on-premise deployment options for government and firm clients.
Generalists until we can justify specialists. The same person may own RAG evaluation today and customer success tomorrow. In early-stage small-market companies, overlap is not a failure mode — it is the only way to move fast.
A note for people who want to apply
If you read this and thought "that sounds like me," we would like to hear from you — even if your profile does not match a traditional job posting exactly.
We have had candidates tell us they are interested but feel they do not have "legal" on their résumé. That is not the primary filter. The filter is whether you care enough about accuracy, about Mongolian context, and about building something that professionals will rely on.
We also hire when we are not posting publicly. If you want to build legal AI infrastructure in Mongolia, write to hello@sodura.ai with what you do and what you want to build.
Closing
Building legal AI in a small market requires a different kind of team. Not just because the market is small, but because the work demands both technical excellence and deep domain care.
We are hiring for the next phase: more AI capacity, more legal accuracy, more products, more markets. But we will not hire bodies. We will hire people who want to build with us — and who recognize that what we are building matters.
If that is you, we are already looking.