AI Engineer

About the job

Project Overview Build AI‑driven Sales Intelligence MVP, CRM‑like UI Users upload Excel/Word discovery docs → auto‑generated digital questionnaire AI matches transcript answers to predefined questions, assigns confidence scores Visualise questionnaire sections (business overview, goals, etc.) on a sales‑stage board Core Requirements Spreadsheet/Doc upload → parse question hierarchy, answer types (checkbox, dropdown) Auto‑create online questionnaire with answer fields Ingest Zoom/Teams recordings, pull transcripts, map answers to questions Confidence scoring (high/medium/low) displayed with color coding Ability to add new question templates per client (20‑200 questions) AI Insight Layer LLM to scan transcripts, extract answers, match to template questions Retrieval‑augmented generation to feed templates into model for better matching Confidence metric attached to each extracted answer Open to Gemini (cheaper) or Claude (current PoC) depending on accuracy needs Integration Points Zoom & Microsoft Teams APIs for recording & transcript retrieval Optional downstream ingest from Gong or other sales enablement platforms (transcript only) AWS S3 for file storage, Lambda for processing, API Gateway for routing Future hook for CRM integrations (optional) Visualization & Dashboard Interactive React dashboard:filter by deal stage, product line, customer segment Drill‑down into individual answers, view confidence scores, export reports Potential static vs. interactive debate; leaning toward interactive with Text‑to‑image for org‑chart/stakeholder maps – consider LLM‑generated specs + D3 rendering Tech Stack & Implementation Details Frontend:React + TypeScript, dynamic form rendering Backend:(Express/NestJS), API layer Parsing:xlsx for Excel, docx or LLM‑assisted parsing for Word Database:DynamoDB (flexible) or PostgreSQL (relational) for questionnaire data AI:Gemini or Claude via API, RAG for template‑driven extraction Storage/Compute:AWS S3, Lambda, API Gateway, Cognito for auth Visualization libs:, for charts & org‑chart generation Open Questions / Decisions Preferred LLM provider (cost vs. accuracy) for production Dashboard interactivity level (static snapshots vs. drill‑down UI) Exact file formats to support initially (Excel only or also Google Sheets/Word) How many confidence levels to display (color-coded three tiers) Timeline for Phase 1 (upload + questionnaire) vs. Phase 2 (insight mapping & dashboard) Contract duration of 1 to 3 months. with 30 hours per week. Mandatory skills: LLM integration, AI Model Development, API, Artificial Intelligence, React, Amazon Web Services, Web Application, AI App Development, Java.

Requirements

• LLM integration
• AI Model Development
• API
• Artificial Intelligence
• React
• Amazon Web Services

Preferred Technologies

• LLM integration
• AI Model Development
• API
• Artificial Intelligence
• React
• Amazon Web Services