AI Engineer (Founding / Full-stack GenAI)

Viraya SKAI Pvt Ltd

Location: Primarily work from home; candidates must be based in Delhi NCR or Kolkata and be available for occasional client visits.

Employment: Full-time

Reports to: Founder (A G)

Experience: 3–6 years in software/ML engineering, including at least 1.5–2 years building LLM applications used by real users or clients.

Compensation: Competitive compensation, commensurate with skills, experience, and demonstrated ability.

ABOUT THE ROLE

We are hiring a hands-on AI Engineer to build and ship LLM-powered products and automation solutions for education and business clients, from first prototype to production.

ABOUT US

Viraya SKAI Pvt Ltd builds AI products and automation solutions for educational institutions and businesses in India. Client services work funds and de-risks our product roadmap, and our products are where we build long-term intellectual property.

Our roadmap spans prompts, retrieval, voice, evaluation, integrations and cost control. We need one strong engineer who can own the full stack, ship fast, and turn one-off builds into reusable components. As our first engineering hire, you will work with freelancers and later hire and lead 1–2 engineers.

WHAT YOU WILL WORK ON

- Conversational AI: text and voice agents that hold structured, multi-turn conversations in English and Indian languages.
- Evaluation and feedback systems: scoring responses or outputs against rubrics and generating useful, personalised feedback.
- AI learning content: applied-AI training material, hands-on labs and auto-graded exercises.
- Workflow automation: document processing, data extraction, reporting and follow-up workflows for business teams.
- Integrations and dashboards: connecting AI services to business tools and messaging channels,



and surfacing results to clients.

KEY RESPONSIBILITIES
- Build LLM applications end to end: prompt design, RAG pipelines, tool-calling agents, backend APIs and simple front-ends.
- Build real-time voice agents using speech-to-text, LLMs, text-to-speech and telephony, with low latency and support for Indian languages and accents.
- Design test sets, rubrics, LLM-as-judge scoring and regression checks so quality is measured.
- Integrate with client systems such as accounting software, CRMs, spreadsheets, email and messaging platforms.
- Deploy and run what you build: cloud hosting, logging, monitoring and cost tracking per user or transaction.
- Handle data responsibly, including user privacy, consent and compliance with India's DPDP Act.
- Create training content, documentation and demos; run occasional workshops or client sessions.
- Scope requirements directly with clients, estimate effort and demonstrate working prototypes.
- Turn repeated client work into reusable modules and templates.
- Choose closed and open-source models and vendors pragmatically based on cost, latency and quality.

MUST-HAVE SKILLS AND EXPERIENCE
- 3–6 years of software or ML engineering experience, including at least 1.5–2 years shipping LLM-based products used by real users or clients.
- Strong Python skills (FastAPI/Flask, async)



and working JavaScript/TypeScript skills for simple front-ends or dashboards.
- Prompt engineering, structured outputs/JSON, function/tool calling, RAG with a vector database such as pgvector, Qdrant or Pinecone, and agent frameworks such as LangChain, LlamaIndex, LangGraph or equivalent.
- At least one voice bot built using STT/TTS tools such as Deepgram, Whisper, Google, ElevenLabs or Sarvam, and telephony such as Twilio, Exotel, Plivo, or platforms like Vapi, Retell or LiveKit.
- Experience extracting structured data from PDFs, invoices, bank statements or forms using OCR and LLMs.
- REST APIs, webhooks, WhatsApp Business API and workflow tools such as n8n, Make or Zapier.
- Docker, one major cloud platform (AWS, GCP or Azure), Postgres, basic CI/CD and monitoring.
- Experience building test sets and measuring LLM output quality.
- Fluent English and Hindi; Bengali is a robust plus for Kolkata-based candidates.
- Comfortable being the only engineer: scoping, building, deploying and talking to customers.

PORTFOLIO REQUIRED Include GitHub links, live demos, or a short write-up of 2–3 LLM or voice projects you built, describing what happened when real users used them.

NICE TO HAVE

- Teaching, mentoring or content creation experience, including workshops, YouTube, bootcamp teaching assistance or guest lectures.
- EdTech experience with learning platforms or assessment design.
- Exposure to finance or marketing workflows in small and mid-sized businesses.
- Indic language models and tools such as Sarvam, AI4Bharat or Bhashini.
- Fine-tuning or running open-source models for cost reduction.
- Startup or freelance experience delivering client projects on fixed timelines.

Work Location: Hybrid remote in Kolkata, West Bengal

📌 AI Engineer (Founding / Full-stack GenAI) (Kolkata)
🏢 Kedia Hardware
📍 Kolkata