Adeel Ghazanfar

Product Manager at Red Buffer | Scaling AI workflows that cut manual overhead

Based in: Islamabad, PakistanCurrently: Product Manager - AI & Digital Products, Red BufferExperience: 4 years

Hire me

Product Manager specializing in AI-driven workflows and revenue intelligence. At Red Buffer, I led the development of an LLM-based extraction system that reduced manual review time from 3 weeks to 6 hours with 93% accuracy. Expert in RAG, vector databases, and multi-model AI evaluation with a focus on bridging complex industry logic with scalable AI architectures.

Experience

Product Manager - AI & Digital Products

Red BufferworkSep 2025 – Present

Summary

Noble Royalties Project: Led product development of a data-intensive oil & gas revenue intelligence platform, translating complex deal, evaluation, closing, onboarding, asset, GIS and revenue workflows into product requirements and technical specifications. TerraCAD Project: Owned ~75% of the product's AI evaluation, regression and testing strategy for a prompt-to-AutoCAD workflow.

What I did

  • Worked directly with engineering teams on LLM-based document extraction, RAG and vector database architecture, defining where AI, retrieval and deterministic processing should be used across the product.
  • Defined workflows for extracting structured information from unstructured documents and connecting extracted data to Deals, Units, Tracts, Wells, ownership and revenue entities.
  • Contributed to AI-driven validation and anomaly detection across revenue consistency, structural data and ownership interests, with emphasis on traceability and user trust.
  • Worked across Next.js, NestJS, Python pipelines, PostgreSQL/RDS, S3, ECS/Fargate and PostGIS/GIS architecture, translating technical constraints into product decisions.
  • Owned requirements, backlog prioritization, sprint planning, UX workflows, technical documentation and stakeholder communication across engineering, UX, QA, DevOps and client teams.
  • Evolved the underlying product/data schema across multiple iterations while managing dependencies between business requirements, data models and engineering implementation.
  • Designed evaluation approaches to assess generated CAD outputs against expected results and identify regressions across model/prompt changes.
  • Worked across product requirements, AI behavior, evaluation criteria and engineering implementation.

Results

  • Reduced document processing time for 250+ PDFs from 3 weeks to just 6 hours (2h extraction + 4h review).
  • Evaluated Claude, GPT, and Gemini models using custom metrics (WER, CER, match rate) to achieve ~93% extraction coverage.
  • 93% extraction coverage using Gemini-based LLM workflows.
  • 90%+ reduction in manual review time for complex oil & gas documentation.
LLMRAGVector DatabaseNext.jsNestJSPythonPostgreSQLRDSS3ECS/FargatePostGISGISAutoCAD

Product Manager

PookiDevs TechnologiesworkMay 2024 – Sep 2025
increase in paid user conversion.
19%increase in paid user conversion.
active users.
3,000+active users.

Summary

GlobalChat - Prompt to content (https://globalchat.app/): Led product development of a privacy-focused multi-model AI workspace integrating GPT, Claude, Gemini and Grok through a unified interface.

What I did

  • Worked across the frontend and AI backend using Next.js, Vercel AI SDK, Node.js and LangChain, including model routing and conversational workflows.
  • Defined product workflows around model selection, context management, AI interactions, subscriptions and user experience.
  • Contributed to architecture integrating PostgreSQL/NeonDB, Stripe and Vercel.

Results

  • Revamped onboarding workflows and revised pricing structure based on GA4 funnel analysis, driving a 19% increase in paid user conversion.
  • Scaled the platform to over 3,000 active users through product-led growth initiatives.
GPTClaudeGeminiGrokNext.jsVercel AI SDKNode.jsLangChainPostgreSQLNeonDBStripeVercelGA4Funnel Analysis

Product Manager

Codistan VenturesworkMay 2023 – Apr 2024

Summary

Managed a portfolio of 7+ concurrent projects across the sports, entertainment, and B2B industries, ensuring alignment between client vision and technical delivery.

What I did

  • Managed a portfolio of 7+ concurrent projects, overseeing resource alignment, change requests, and milestone closures.
  • Ensured project and client vision remained intact throughout the development lifecycle by bridging the gap between stakeholders and engineering teams.
  • Coordinated across multiple workstreams to ensure on-time delivery and adherence to project roadmaps.

Results

  • Successfully managed and delivered 7+ concurrent projects simultaneously, ensuring all milestones were met and client visions were maintained.

Skills

Technical

Agile/Scrum
Sprint Planning
AI Workflows
RAG
Prompt Engineering
SlackSlack
NotionNotion
MS Office
Product Discovery
JIRAJIRA
LLMs
Product Roadmaps
PRDs
User Stories
Acceptance Criteria
Backlog Management
Wireframing
Vector Databases
AI Evaluation
LangChain
Vercel AI SDKVercel AI SDK
Multi-Model AI
Data Extraction
UX
Next.jsNext.js
Node.jsNode.js
REST APIs
PostgreSQLPostgreSQL
Git/GitLabGit/GitLab
KPI Tracking
FigmaFigma
User Flows
Information Architecture
Dashboard Design
Usability Refinement
ClickUp
Confluence
GA4
S3
NestJS
Funnel Analysis
RDS
AWS
PythonPython
Anomaly Detection
Cohort Analysis
Retention Curves
A/B Testing
ECS/Fargate
PostGIS

Areas of expertise

Stakeholder Management
Data-driven Decisions
UX Team Leadership

Projects

TerraCAD Project

Product Managerprofessional
Owned ~75% of the product's AI evaluation, regression and testing strategy for a prompt-to-AutoCAD workflow. Designed evaluation approaches to assess generated CAD outputs against expected results and identify regressions across model/prompt changes. Worked across product requirements, AI behavior, evaluation criteria and engineering implementation.
AutoCADAI EvaluationRegression Testing

GlobalChat - Prompt to content

Product Managerprofessional
Led product development of a privacy-focused multi-model AI workspace integrating GPT, Claude, Gemini and Grok through a unified interface. Worked across the frontend and AI backend using Next.js, Vercel AI SDK, Node.js and LangChain, including model routing and conversational workflows. Defined product workflows around model selection, context management, AI interactions, subscriptions and user experience. Contributed to architecture integrating PostgreSQL/NeonDB, Stripe and Vercel.
Next.jsVercel AI SDKNode.jsLangChainPostgreSQLNeonDBStripeVercelGPTClaudeGeminiGrok
Visit the site

Noble Royalties Project

Product Managerprofessional
Led product development of a data-intensive oil & gas revenue intelligence platform, translating complex deal, evaluation, closing, onboarding, asset, GIS and revenue workflows into product requirements and technical specifications. Worked directly with engineering teams on LLM-based document extraction, RAG and vector database architecture. Defined workflows for extracting structured information from unstructured documents and connecting extracted data to Deals, Units, Tracts, Wells, ownership and revenue entities. Contributed to AI-driven validation and anomaly detection across revenue consistency, structural data and ownership interests.
Next.jsNestJSPythonPostgreSQLRDSS3ECS/FargatePostGISGISLLMRAGVector Database

Education

Bachelor of Science

Bahria UniversityComputer Science2018 – 2022
Hire AdeelReach out about a role, a contract or a conversation.For recruiters

Adeel's twin is AI, it can make mistakes.

More people on Kanvis

Browse more peopleCreate your own Kanvis