Back to all roles

AI Platform

/

Senior leadership

Staff LLM Engineer — Foundation Models

LLM Training & Serving

Lead Full-Time Remote (India preferred) / Guwahati Permanent

Compensation

INR 50 - 80 LPA + equity.

Engagement

Full-Time

Permanent role. Full-time commitment. Remote-first, with periodic in-person off-sites.

Scope of role

Set direction within your domain. Build and mentor a team. Own outcomes at the function level.

01 — The role

Why this role exists at EduRankAI

Own the LLM training stack. Pre-train, supervised fine-tune, RLHF/DPO, evaluation, release. You will lead the recipes that produce EduRankAI's foundation models and you will be the engineer the rest of the platform calls when training stalls at 3am. This is a hands-on Staff role. You will write code, run trainings, debug NaNs, profile kernels. You will also mentor a small team of mid/senior LLM engineers and shape our training methodology over multiple model generations.

02 — The work

What you will own

  • 01 Own the pre-training pipeline for 1B → 70B parameter transformer models.
  • 02 Lead SFT and preference-optimisation runs (RLHF, DPO, KTO, IPO as the field moves).
  • 03 Design the eval harness — Indic reasoning, factuality, safety, refusal calibration.
  • 04 Drive inference optimisation jointly with the serving team (KV cache, quantisation, batching).
  • 05 Mentor the LLM engineers; bring at least one to Staff readiness over 24 months.
  • 06 Author public technical reports where the work merits it.

03 — The expertise

What we look for

Trained 7B+ parameter transformers from scratchPyTorch + JAX + CUDA at depthDistributed training (FSDP / Megatron / DeepSpeed)RLHF / DPO implementation experiencevLLM / SGLang / TensorRT-LLMTokenizer + data pipeline expertise

04 — The bar

Who thrives here

  • You have led at least one foundation-model training run at 7B+ parameters on a cluster you owned or operated.
  • You can debug a loss curve and propose a hypothesis within an hour.
  • You have published or open-sourced ML/LLM work the field has cited or used.
  • You can read the latest frontier-paper repo, identify why the training method is fragile, and write a one-pager about how to harden it.
  • You write your own evals.

05 — How we work

The EduRankAI environment

Remote-first, async-first

Work from anywhere. We optimise for deep work, not face time. Periodic in-person off-sites for the full-time team.

High autonomy, high standards

We hire adults and trust them. You will be expected to set your own goals, communicate clearly, and ship.

Builders, not bureaucrats

We optimise for clarity over process. Make the call, ship the work, write up what you learned.

Bharat-built, globally ambitious

We are an Indian frontier AI lab. We build for India first and the world second — in that order.

06 — Hiring process

What to expect after you apply

  1. 01

    Application review

    Every application is read personally within five business days. We respond either way.

  2. 02

    Take-home or live exercise

    Role-specific. Time-boxed. Real problems we are actually working on, not invented puzzles.

  3. 03

    Conversations

    Deep technical and values conversations with the team you would join. No trick questions. No panel ambushes.

  4. 04

    Offer or honest no

    If yes: digital offer letter, signed in-portal, transparent terms. If no: written feedback if you want it.

Before you start

What we will collect. What it costs. What we will not do with it.

Application fee

CHF 100

Lead tier

We will collect

  • Name, email, phone — Account + application updates. No marketing.
  • Resume / portfolio link — Human review of your work.
  • Date + place of birth — Identity verification only.
  • Your written responses — Selection rubric. Read by humans.
  • Government ID (later) — Anti-fraud at offer / interview stage. Not at signup.

We will never

  • Sell your data
  • Share with third-party recruiters
  • Use for advertising
  • Train models on it
  • Send marketing email

Our situation

EduRankAI is a small, independent organization building long-term capabilities in educational intelligence, advanced AI systems, and research infrastructure. We take no advertiser money, no donations with strings attached, and no investor pressure on hiring decisions. The small per-application fee covers the real cost of processing your application — human review, identity verification, infrastructure, reviewer time. It buys us the right to be honest. Genuine financial hardship? Request a fee waiver inside the application — reviewed individually within 5 business days, with no record in your file and no second-class treatment of waiver-granted applications.

Full transparency policy Why we charge a fee Questions? Email us

Ready to apply?

We read every application personally. If you are the right person for this role — regardless of pedigree, background, or where you are based — you will hear back from us within five business days.

Lead Full-Time

Staff LLM Engineer — Foundation Models

Apply →