E-Gmat
Website:
e-gmat.com
Job details:
About PRiSM Learning / e-GMAT
PRiSM Learning is the company behind e-GMAT, the brand that pioneered online GMAT prep and remains the #1 prep company in the category: 3,000+ reviews on GMAT Club at a 4.7 rating, more than three times the reviews of any competitor.
Founded by alumni of IIT Delhi and BITS Pilani, e-GMAT was built at the intersection of learning science, pedagogy, data and technology. Our co-founder Payal Tandon, the world's #1 GMAT instructor, has guided tens of thousands of students to their target scores. Rather than scaling through more instructors, we encoded her pedagogy into AI-powered products. Over the last five years, e-GMAT has sent more students to M7 business schools and ISB than any other prep company in the world.
We serve 10,000+ paying students a year with a team of 10. Our vision is to reach 5,000,000+ students over the next four to five years, across every competitive exam: GMAT, GRE, SAT, ACT, IELTS, GATE and more.
Build the technology that carries the teaching
e-GMAT pioneered self-paced learning, and then did the much harder thing: made it work better than a classroom. Not more convenient - more effective. A student working alone with our product learns more than they would in a room with a teacher, and more than they would with any other course available to them.
That is not an accident of content quality. It is the result of building technology for a pedagogical purpose - every system designed around how people actually learn, and judged by whether they learned. Our co-founder Payal Tandon is the world's #1 GMAT instructor; rather than scale by hiring instructors like her, we encoded her pedagogy into the product. Over the last five years that approach has sent more students to M7 business schools and ISB than any other prep company in the world.
You are here to build the next layer of that, for a million paying students.
What you are actually amplifying
Every other hire we make adds one person's output. This role multiplies everyone else's - and what it multiplies is not any individual's hours. It is the pedagogy itself.
Our subject matter experts design how something should be taught. Our product encodes it. The distance between those two things - a method that works in one expert's hands, and a system that delivers it to a million students who will never meet them - is the distance you close.
The same applies inside the company. A capability you build for one of our teams should double what a person in that role can do. Not trim ten minutes off a task: change what the job consists of. That kind of impact is rare, and it is the standard here.
A concrete example. A member of our customer success team interviews a student who just hit their target score. Today that becomes one write-up. It should become a social carousel, an Instagram story, a short and an internal debrief - so that team publishes like a marketing team without hiring one. That is a single system for a single team, and there are a dozen more waiting across the company.
What you will own
Solutions for studentsYou build what students use, against our strategic plan for the year. The bar is ours and it does not move: a feature that exists and is used is not a success. Students have to be measurably better off, against a comparison group, or we have not done anything.
The pedagogy is designed by our subject experts, not by you. You decide how it gets built, what it should cost to run, and how we know it worked.
Capabilities that double what our teams can doGo to a team. Understand the work as it is actually performed, not as the org chart describes it. Then build the capability that changes what they are able to do.
This is not handing someone a tool for the process they already run - that leaves a person with one more thing to operate and no faster. The test is whether the team can now do something it previously needed a specialist for: someone it does not employ and, because of you, will not need to.
Processes re-designed, not decoratedWhere the work above raises what a person can do, this removes work that should not need a person at all. Two hundred forum posts a week. Reporting pulled by hand each month. Read-outs a founder currently assembles.
You decide what is worth automating, and you prove the hours genuinely went away rather than moving somewhere less visible - that the exception rate does not eat the saving, and that what an agent costs to run stays below what it replaced.
A fleet of thirty-plus agents that stays current
You own an evaluation for every one of them. Not a sample - every agent.
That discipline is what makes everything else safe to change. It is what lets you adopt a better model the week it ships, cut what an agent costs to run, and collapse three steps into one, without discovering later that quality quietly fell. Without it a fleet this size decays while still appearing to work. It is one of the most important things about this role, and it is routinely the thing engineers skip.
Why you will want this
- You will work with the best tools, not a procurement committee's version of them. Everyone here works with Claude and Codex daily - not just the harnesses but the tooling built around them. When something better ships, we use it.
- Our founders are AI-native. They build with these tools themselves. You will not spend your first six months explaining why an agent is worth trying.
- You will have room to experiment. The technology changes every few months, which means the only honest way to know what works is to try it. That freedom is real here, and it is funded - we would rather you spent money finding out than guessed cheaply.
- You will see your impact. Not in a quarterly deck. In a team that now does something it could not do last month, and in students whose results changed.
The experience and judgment you bring
You have built systems other people depended on, and you can explain one end to end: the problem, who you talked to, what success had to mean, the architecture you chose and the ones you rejected, which components needed a model and which needed code, what broke first, and how you found out.
Your strengths combine:
- System design from a real problem. You start with the problem and the stakeholders, establish what success actually means, and design from there - not from the tool you wanted to use.
- Judgment about where a model belongs. You know which parts of a system should be LLM-driven and which should not, and you can defend both calls.
- Prompts that are simple and clear. Simplicity is the standard, not cleverness.
- Fluency with the model landscape. What each model can do, what it costs, what it is genuinely suited for - and you keep up, because the answer changes every few months.
- Evaluation as a habit, not a phase. You do not consider something finished until you can tell whether a change made it worse.
- Building large systems in pieces. Subcomponents built and tested independently, and real skill at debugging what you assemble from them.
- A process for extracting what people need. Internal customers rarely describe their problem accurately. You have a way of getting to the real one, and you can describe that way.
- Process flowcharting. You can draw how work actually moves before you change it.
You should want the multiplier job specifically. If your satisfaction comes from owning one product surface deeply, this will frustrate you - the work is broad, it lands inside other people's teams, and much of your best output will show up as someone else's results.
Compensation
- Base: ₹30 lakhs per year.
- A serious performance bonus.
- Equity in the business. You are building something you will own a share of.
PRISM Learning is an equal opportunity employer. We evaluate candidates on capability and potential, not background or pedigree.
Click on Apply to know more.
