Afeef Ahmed

Architected a ChatGPT-like RAG chatbot using LangChain, GPT-4o, and Pinecone vector database for IPAM (IP Address Management) -
enabling natural language queries for IP conflicts, subnet utilization, and device assignments, reducing network engineer ticket resolution
time by 60%.
Architected a ChatGPT-like RAG chatbot using LangChain, GPT-4o, and Pinecone vector database for IPAM - enabling natural language
queries for IP conflicts, subnet utilization, and device assignments, reducing network engineer ticket resolution time by 60%.
Processed Invalid IPAM records using LLMs - enabled SPOC owners to claim, validate, and adjust their IP assignments, cutting manual
reconciliation time by 70%.
Fine-tuned LLMs for IP address suggestions and subnet conflict detection - cut provisioning time from days to under 2 hours.
Developed production-grade REST APIs (FastAPI, .NET Core) to serve LLM-powered IPAM recommendations, semantic search, and
anomaly predictions - handling 10, 000+ daily requests with sub-200ms latency and 99.9% uptime on Azure Kubernetes.
Implemented MLOps + LLMOps pipelines using Docker, GitHub Actions, MLflow, and Azure DevOps - automated model retraining,
versioning, and deployment, increasing model update frequency by 70% while maintaining

Developed AI pipelines to convert unstructured data into structured data for National Cybercrime Forensics Lab. Enabled real-time
•
•
intelligence and faster strategic decisions.
Performed advanced analytics on restricted marketplace data, including sentiment analysis and topic modeling, leading to early threat
detection and 30% faster insight generation.
Managed secure web scraping and data preprocessing, designing predictive workflows and dashboards with Elasticsearch and Kibana
that improved monitoring efficiency by 40%.
Authored an NLP research paper on automated information extraction, contributing to the labs AI research pipeline and next-generation
threat detection models.

Architect and optimize Apache Spark jobs (batch & streaming) for large-scale data transformations, achieving high throughput and low
•
•
latency for critical public health dashboards.
Design, develop, and maintain ETL pipelines and distributed data processing systems for healthcare intelligence, processing terabytes of
real-time Polio monitoring and patient care data.
Architected scalable RAG solutions using GPT and Llama models with vector databases (Pinecone, Faiss), achieving 95% retrieval
accuracy and reducing manual search time by 60% through intelligent web interfaces.
Deploy and manage production-grade solutions on AWS (EC2, S3, EMR/Glue, IAM) using containerized microservices, CI/CD pipelines,
and automated versioning, ensuring 99.9% uptime.
Directed AI architecture for healthcare intelligence initiatives, including Health Guard AI and enterprise RAG systems, strengthening
decision-making in public health and polio monitoring programs.

Developed and optimized large-scale ETL pipelines using Python, Spark, and SQL to ingest, process, and analyze petabytes of
•
•
unstructured social media and dark web data.
Designed automated data-gathering frameworks using custom scrapers (Selenium, Scrapy) to collect and process large-scale social
media datasets, powering NLP and LLM models for unstructured text analysis.
Architected and optimized robust data pipelines (Python, SQL, Spark, Elasticsearch) for user activity analysis, behavioral profiling, and
trend detection across multiple social media channels.
Conducted R&D on NLP techniques including sentiment analysis, entity recognition, and pattern discovery to detect emerging cybercrime
trends and uncover hidden connections in unstructured data.

Designed algorithms and predictive models to extract insights and analyze trends from complex datasets, resolving critical business
•
•
•
problems through statistical analysis.
Built advanced web automation frameworks using Selenium, Scrapy, TOR, and proxy networks with custom CAPTCHA-solving algorithms,
enabling large-scale, real-time data collection from web forums.
Developed Flask-based RESTful APIs to structure unstructured forum data, powering AI pipelines for NLP, entity extraction, and
automated cyber intelligence.
Co-led a team in developing an end-to-end data visualization tool, enhancing strategic planning and decision-making capabilities.

Accelerated strategic outcomes by 25% through statistical analysis of 50+ GB of data using Python, generating actionable business
•
•
• insights for a consumer goods trading company.
Designed interactive dashboards and reports using Power BI and Tableau, enabling real-time performance tracking and KPI monitoring for
management.
Executed data cleaning and preprocessing to ensure high-quality datasets, improving trend forecasting accuracy and reporting reliability.