zenskar
Website:
zenskar.com
Job details:
Description
About this role :
As a Senior GenAI Engineer at Zenskar, you will own the AI layer of our product - building the features that make Zenskar intelligent. This is not a research role and not a prompt-engineering role. You will build production AI systems that enterprise clients depend on, which means reliability, observability, and rigorous evals matter as much as the AI capability itself. You own the full vertical - the model, the pipeline, and the UI.
- Build and own CS Copilot a real-time assistant for customer success teams, spanning STT pipelines, live transcription, and LLM-powered suggestions
- Build LLM-powered document understanding features extracting structured, reliable data from unstructured enterprise documents
- Own AI feature UIs end-to-end you build the interface, not just the model integration layer
- Design and maintain an eval framework define what 'working' means for each AI feature and catch regressions before users do
- Drive model selection and integration decisions choosing the right provider and approach for each use case, managing latency and cost
- Own AI platform reliability - observability, fallback behaviour, and graceful degradation when models fail
- Work closely with product, customer success, and the full-stack engineer - AI features only matter if they are usable and trusted by real users
The Impact You'll Make
- You will define what AI means at Zenskar - the features you ship will be the most visible and differentiated parts of the product
- CS Copilot, if done well, changes how enterprise customer success teams operate every single day - this is a high-stakes, high-visibility surface
- You will establish the engineering culture around AI reliability at Zenskar - evals, observability, and disciplined iteration
- Your work will directly accelerate enterprise deals - AI features are increasingly a buying criterion for our clients
- You will be the person who brings engineering rigour to a domain where most companies ship demos and call it a feature
Key Qualifications
Must have :
- 4 - 6 years of total software development experience, with at least 1- 2 years actively building and shipping AI/LLM-powered features in production
- CS degree or equivalent - strong engineering fundamentals; this role requires a builder, not a researcher
- Can describe the hard AI problems they have solved - not which models they used, but what broke, what they learned, and what changed as a result
- LLM pipeline engineering - chaining, tool use, structured outputs, prompt management, handling failure modes in production
- Prompt lifecycle management : Treats prompts as versioned, testable artefacts with a system for managing changes in production; not ad-hoc strings
- RAG fundamentals : Chunking strategies, retrieval quality, embedding models, and the judgment to know when RAG is the right answer
- Eval discipline : Has designed their own eval sets, knows how to measure AI feature quality and catch regressions when models change
- Latency and cost thinking : Knows when to use a smaller model, when to cache, and when streaming matters
- Responsible AI and data privacy awareness : Understands the implications of passing enterprise data through third-party models; designs pipelines with this in mind from day one
- Product-quality frontend : Can own AI feature UIs end-to-end in React or equivalent
- Strong backend engineering : Can own the full feature without needing a backend engineer to make AI code production-ready
Good To Have
- Agentic systems - multi-step agents, tool orchestration (LangGraph, CrewAI, or custom), long-running workflows
- Voice and multimodal pipelines : STT (Deepgram, Whisper), real-time audio processing, WebSocket streaming
- Fine-tuning experience : LoRA/QLoRA on open-source models; signals understanding of model internals beyond the API surface
- Open-source model deployment : vLLM, Ollama, running models on own infrastructure
- B2B SaaS / enterprise AI experience : Building AI for workflows where correctness and auditability matter
- AI-assisted development : Comfortable building AI tools with AI tools; uses Cursor, Copilot, or similar as a genuine force multiplier
- Experience with financial systems, billing platforms, or fintech applications
- Knowledge of SaaS business models and compliance frameworks
- API design and integration experience
- Prior experience working at a startup
- Not taking yourself too seriously
What Drives You
- You have been embarrassed by an AI feature failing in production - and you fixed it systematically, not with a workaround
- You think about evals before you think about model choice - reliability is the product
- You apply the same engineering rigor to probabilistic systems that you would to deterministic ones
- You own AI features end-to-end - the model, the pipeline, the UI, and the fallback behaviour
- You find it unsatisfying to ship a demo - you want to ship something that holds up under real enterprise usage
- The AI landscape changes weekly and you find that energizing, not exhausting - you are genuinely curious and stay ahead of it
Location :
- Hybrid : 2 days per week
- Office Location : Indiranagar, Bengaluru.
- Address : 3rd Floor, A wing No 1, Carlton Towers, HAL Old Airport Rd, HAL 2nd Stage, Indiranagar, Bengaluru, Karnataka 560008.
(ref:hirist.tech)
Click on Apply to know more.