Stories

Your pitch, six stories that answer most behavioral questions, and the tough questions your resume invites.

Your pitch

Full pitch, about 90 seconds

Starting draft

I’m a Senior Generative AI Engineer at EPAM, where I lead the architecture and delivery of enterprise GenAI platforms. Right now that’s a multimodal hybrid RAG system on AWS Bedrock and OpenSearch for a Tier-1 US financial client, including the LLMOps layer: evaluation, guardrails, PII protection and cost controls.

I came into AI from engineering. I started as a full-stack Django developer, then spent eighteen months as a data engineer at TCS building GCP pipelines, where I cut cloud costs by 40%. In 2022 I moved into GenAI at Cognizant and took semantic search and RAG from prototypes to production. At PwC I architected multi-agent services on AWS that cut risk-assessment turnaround by 70% and workflow latency by 68%.

Outside work I build what I’d want as an architect: a multi-tenant LLM gateway, a hybrid RAG platform with document-level access control, and EvalSeed, an evaluation-dataset library on PyPI. I’ve also interviewed more than fifty GenAI engineers at EPAM, which shapes how I think about engineering standards.

Now I’m looking for a role where I own GenAI architecture end to end, and [why this company].

30-second version

Starting draft

I’m a Senior GenAI Engineer at EPAM with six years across full-stack, data engineering and GenAI. I design and ship production RAG and multi-agent systems on AWS and GCP, most recently a multimodal hybrid RAG platform for a US financial client, along with the evaluation, guardrails and cost controls that keep them reliable. I’m looking to own GenAI architecture end to end.

Six stories

0 of 6 drafted
A technical win you droveTell me about the achievement you’re proudest of.Leading without authorityTell me about a time you got other teams to adopt your approach.A production failureTell me about a time something you built failed in production.Pushing back on a stakeholderTell me about a time you disagreed with a client or stakeholder.Raising the barHow have you helped other engineers grow?A decision you got wrongTell me about a decision you got wrong.

Tough questions

Say
  • Own it calmly. Each move was a deliberate step: full-stack, then data engineering, then GenAI, then architecture on AWS, then enterprise GenAI at EPAM.
  • Show the thread: every move added a layer you now use as an architect.
  • Say what you want next and why this role gives you that for the long term.
Avoid

Blaming past employers, or making it only about pay.

Say
  • Frame it as a pull, not a push: broader architecture ownership, a product or domain you care about, or a platform role.
  • Acknowledge what EPAM gave you.
Avoid

Criticizing your current client or manager.

Say
  • Depth over years: production systems shipped, decisions you owned, and standards you set as an interviewer.
  • Offer evidence: “Give me your hardest design question.”
Avoid

Arguing that years don’t matter.

Say
  • Research the band first: AmbitionBox, Levels.fyi and peers in similar roles.
  • Give a range anchored on the role’s scope and the market, not just your current pay.
  • If they ask for your current pay, answer honestly, then bring the conversation back to the role.
Avoid

Naming a number before you know the role’s level.

Say
  • Know the exact date, whether it can be shortened, and any buyout option. Say it plainly.
Avoid

Being vague. Recruiters drop candidates whose availability is unclear.

Say
  • A real gap with a plan already underway, such as GPU serving and inference tuning, which you’re closing by self-hosting a model with vLLM.
Avoid

“I work too hard.”

Say
  • Name specific sources, and one thing you built or changed because of them.
Avoid

“I read a lot on LinkedIn.”

Say
  • Draw on your 50+ interviews: production judgment over buzzwords, how candidates reason about trade-offs, and whether they measure results.
Avoid

A generic answer. This is where you show bar-raiser judgment.

Questions to ask them

What should someone in this role have delivered in the first 90 days?
Which GenAI systems are in production today, and what’s the hardest open problem?
How do you evaluate quality before and after a release?
Who makes architecture decisions, and how are they reviewed?
How is time split between hands-on building and design or client work?
A year from now, what would make you say this hire was a great decision?