Skip to content
View iamalimaybe's full-sized avatar
🎯
Focusing
🎯
Focusing
2 followers · 0 following
  • EvReady Pakistan

Achievements

Achievement: Pull Sharkx3Achievement: YOLO

Achievements

Achievement: Pull Sharkx3Achievement: YOLO

Block or report iamalimaybe

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
iamalimaybe/README.md

Featured Projects

EVReady Pakistan

A deployed Pakistan-focused EV platform with a public frontend, public backend API, and internal AI recommender service.

The architecture routes AI recommendations through the backend gateway instead of exposing the recommender or local model runtime directly.

Shows: Spring Boot API gateway, React/Vite frontend, PostgreSQL, Liquibase, Docker deployment, internal service integration, async AI recommendation flow, rate limiting, persisted recommendation runs, and local Ollama/Qwen model usage.

Repos: Frontend · Backend · AI Recommender

RAG Document Assistant

A Java/Spring Boot RAG backend with a thin React UI for grounded document question answering, citations, missing-info handling, QA run history, and audit visibility.

Shows: document upload, text extraction, chunking, local embeddings with Ollama, PostgreSQL + pgvector retrieval, prompt context budgeting, structured LLM output parsing, citation validation, retrieved evidence snapshots, and persisted QA runs.

Why it matters: RAG answers are not accepted blindly. Retrieved chunks, cited chunks, prompts, model output, and QA runs are stored so answers can be inspected and improved.

Release: v0.3-thin-react-ui Repo: rag-document-assistant

LLM Evaluation Registry

A Java/Spring Boot backend quality layer for evaluating and comparing LLM workflow behavior across prompt versions, model providers, reusable evaluation cases, individual runs, and queued batch runs.

Shows: prompt versioning, evaluation cases, provider abstraction, Ollama and OpenAI execution paths, raw and parsed model output storage, deterministic scoring, regression comparison, batch execution, cancellation, and review notes.

Why it matters: Prompt and model changes can silently make AI workflows worse. This project makes LLM behavior measurable, comparable, and auditable.

Release: v0.6-openai-provider Repo: llm-evaluation-registry

AI Ticket Triage Service

A Java/Spring Boot backend for structured support ticket triage with validated AI output, PostgreSQL persistence, auditability, and a lightweight React review console.

Shows: local LLM integration through Ollama/Qwen, deterministic analyzer fallback, structured output validation, raw model output storage, confidence-based review routing, review statuses, consistent API errors, Swagger docs, Docker Compose, and CI.

Why it matters: AI output is treated as untrusted until it is parsed, validated, persisted, and routed through review rules.

Release: v0.3-frontend-review-console Repo: ai-ticket-triage-service

Requirements Intelligence Assistant

A local LLM workflow for software requirement analysis where model output must pass schema validation, semantic checks, regression tests, and structured run-report validation.

Shows: trusted context validation, local Ollama/Qwen execution, malformed JSON repair, output normalization, enrichment from trusted context, JSON Schema validation, semantic validation, negative tests, multi-context regression tests, and PASS/FAIL run reports.

Why it matters: The workflow detects unsupported, incomplete, or hallucinated model output before it is accepted.

Release: v0.1-validation-first-local-llm-workflow Repo: requirements-intelligence-assistant

Pinned Loading

  1. ev_ready_back ev_ready_back Public

    Backend API for EVReady Pakistan with vehicle catalogue, charger directory, leads, contact submissions, reviews, feedback, admin moderation, and cataloge management.

    Java

  2. evready-ai-recommender-service evready-ai-recommender-service Public

    Catalog-backed AI recommender service for EVReady Pakistan using trusted vehicle facts, deterministic candidate filtering, structured LLM output, validation, and stored recommendation runs.

    Java

  3. rag-document-assistant rag-document-assistant Public

    Grounded RAG document QA backend with citations, missing-info handling, pgvector retrieval, structured model output validation, and QA run history.

    Java

  4. llm-evaluation-registry llm-evaluation-registry Public

    Backend-led quality layer for AI workflows. Tracks prompts, models, evaluation cases, model runs, validation results, regressions, batch comparisons, and review notes so LLM behavior can be measure…

    Java 1

  5. ai-ticket-triage-service ai-ticket-triage-service Public

    Java/Spring Boot AI backend for structured support ticket triage with validated LLM output.

    Java

  6. requirements-intelligence-assistant requirements-intelligence-assistant Public

    Validation-first local LLM workflow for grounded software requirement analysis using trusted context, schema validation, semantic checks, regression tests, and Ollama.

    Python 2