About Us
SpectraSeek powered by InterspectAI is a cutting-edge AI-powered interview preparation platform, built to help students and job seekers practice real-world, role-specific interviews with instant, personalized feedback. Backed by InterspectAI, our mission is to make every student career-ready and every institution AI-prepared.
The Role
We're looking for a blended Data Scientist + LLM Engineer to help shape the future of intelligent interviewing through multi-agent orchestration, prompt-driven systems, and workflow-optimized AI.
You’ll be at the core of our Spectra engine — fine-tuning prompts, building agent chains, and infusing LLM intelligence into dynamic interviews for hiring, coaching, and assessments.
Key Responsibilities
- Design and implement multi-agent interview orchestration flows using tools like LangGraph, LangChain, AutoGen, or similar.
- Own the prompt engineering lifecycle — design, iterate, test, evaluate.
- Collaborate with product teams to align LLM behavior with real-world interview workflows, feedback loops, and rubric scoring.
- Build and optimize embedding-based RAG pipelines (e.g., FAISS, Weaviate, Pinecone).
- Build LLM evaluation tools for output consistency, reasoning traceability, and scoring metrics.
- Develop test datasets and metrics to benchmark LLM agents (e.g., consistency, tone, hallucination).
- Analyze candidate transcripts, interview sessions, and behavioral feedback to identify new features and areas of model improvement.
Required Skills
- 3–5 years in Data Science or NLP-focused Engineering roles
Strong Python experience in:
- LangGraph, LangChain, LlamaIndex, or similar orchestration tools
- OpenAI, Claude, Gemini, or open-source models (LLama2, Mistral, etc.)
- Prompt templating, prompt routing, multi-step flows
Strong understanding of:
- Vector stores and similarity search
- Embedding models (OpenAI, HuggingFace, etc.)
- Evaluation techniques (BLEU, ROUGE, custom LLM evaluation via GPT)
Familiar with:
- RAG pipelines, structured memory, and retrieval tuning
- Tools like Weights & Biases, MLFlow, or custom tracking
Bonus Points
- Experience designing agent-based workflows (e.g., job interviewer agents, career coach agents, tutor agents).
- Familiarity with LangGraph state management, guardrails, and async agent cycles.
- Comfort working with graph-based execution plans or DAGs for reasoning flows.
- Experience with Voice-based or real-time LLM interaction pipelines.
- Prior work in HRTech, EdTech, or CareerTech is a plus.
Our Stack
- Python, FastAPI, LangGraph, LangChain, OpenAI API
- CockroachDB, Postgres, Redis, AWS (S3, Lambda, Step Functions)
Why Join Us
- Build next-gen agentic AI that impacts hiring, education, and career coaching
- Join a venture-backed startup with real traction and a bold mission
- Collaborate with founders, AI scientists, and startup veterans
- Competitive compensation, equity potential, and flexible location
Type: Full-Time | Contract-to-Hire | Founding Team Consideration
Team: Spectra AI Platform @ InterspectAI
Job Types: Full-time, Contract
Benefits:
- Dental insurance
- Flexible schedule
- Health insurance
- Paid time off
Work Location: Remote
