Skip to content
View syahra712's full-sized avatar

Highlights

  • Pro

Block or report syahra712

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.

Content in all repositories owned by your account will be closed.
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
syahra712/README.md
Syed Ahnaf Raza — Software Engineer & AI Developer



Software Engineer · AI Developer · PMO @ VentureDive . building production RAG, LLM agents & data pipelines.

Portfolio LinkedIn Profile views

🔧 Open Source — Merged into the projects the ecosystem runs on

Not toy PRs. Real bug fixes, reviewed and merged into two of the largest AI projects on GitHub.

Project ⭐ Stars Contribution Status
BerriAI/litellm 57k+ Don't retire a completed batch from cost recovery while output_file_id is lagging Merged
docling-project/docling 65k+ Coerce MLX bfloat16 logprobs to float in VlmPredictionToken Merged

More contributions in flight across the Python / AI-ML ecosystem.


🛠️ Stack

Languages Python TypeScript JavaScript SQL

AI / ML PyTorch TensorFlow LangChain Hugging Face spaCy

Build & Deploy FastAPI React Next.js Docker MongoDB GitHub Actions


🚀 Featured Work

Project What it does
urdukit Zero-dependency NLP toolkit for Urdu — normalization, tokenization, stemming, Roman-Urdu transliteration. No model downloads.
hiddenink Detects & cleans invisible Unicode, homoglyphs and container metadata — per-occurrence decisions on hidden marks vs real content.
Graph-Augmented Research Agent Fuses semantic similarity with co-citation & bibliographic coupling — gap analysis, plagiarism detection, citation formatting.
BridgeSign Real-time ASL recognition — MediaPipe landmarks, 38-class classifier, spell-correction + text-to-speech.
EcoViz Emissions dashboards + a stacked regression benchmark across 8 model families on CO₂ and methane data.
AES-RSA Service FastAPI hybrid AES + RSA encryption for text & files — streaming, compression, digital signatures.

🎖️ Credentials

  • 🏆 Regional Agentic AI Hackathon — Winner
  • 📜 Oracle Certified Generative AI Professional (2025)
  • 📜 Stanford Machine Learning Specialization
  • 📜 Google Project Management Certificate
  • 📜 Microsoft PL-300 (Power BI Data Analyst)

Experience — ML pipelines @ 10Pearls · RAG / LLM systems @ Sybrid


Open to roles in AI/ML & Data Science · Let's build something that ships.

Pinned Loading

  1. DataVisualzationEcoViz DataVisualzationEcoViz Public

    EcoViz — emissions dashboards plus a stacked regression benchmark across 8 model families on CO2 and methane data

    Python 1 1

  2. AIProject-BridgeSign AIProject-BridgeSign Public

    Real-time American Sign Language recognition — MediaPipe hand landmarks, 38-class classifier, spell-correction and text-to-speech

    Python

  3. Graph-Augmented-Intelligent-Research-Agent Graph-Augmented-Intelligent-Research-Agent Public

    Research agent fusing semantic similarity with co-citation and bibliographic coupling — search, gap analysis, plagiarism detection, and citation formatting

    Python

  4. Named-Entity-Recognition-NER-for-Documents Named-Entity-Recognition-NER-for-Documents Public

    Named entity recognition over documents using spaCy and BERT transformers, with a comparison of rule-based and neural approaches

    JavaScript

  5. RAGCHATBOT RAGCHATBOT Public

    Wikipedia-grounded retrieval-augmented chatbot built with Streamlit and Google Gemini

    Python