• I build end‑to‑end analytics and GenAI systems: data pipelines → models → APIs/UI → eval/observability → deployment.
• Tooling I lean on: Python, SQL, scikit‑learn, PyTorch, TensorFlow, LangChain, Hugging Face, FastAPI, Streamlit, MLflow, Docker, AWS.
• What teams get from me: grounded results, clean code, measurable business impact, and docs that drive adoption.

• +40% landing‑page conversions via experimentation and UX audits (Manito Mix)
• 2.4× ROAS with lower CPA on $150K+ annual budget (Manito Mix); 3× ROAS on $200K+ (Kissflow)
• −12% excess inventory, +7% service levels from SKU‑level forecasting (P&G)
• −60% reporting time via automated, client‑ready dashboards

• Earnings Intelligence Copilot (RAG + LangChain)

  • What it does: Answers “What changed vs last quarter?” on 10‑K/10‑Q/8‑K + transcripts with citations; extracts KPIs and flags risk language shifts.
  • Stack: LangChain, FAISS/Qdrant, bge embeddings + reranker, Llama 3/Mistral, FastAPI, Streamlit, RAGAS, MLflow.
  • Outcome focus: Lower analyst triage time; auditable answers; KPI extraction F1 and hallucination rate in reports.

• AP/AR Invoice–PO Reconciliation Agent

  • What it does: OCRs invoices, matches line items to POs, flags duplicates/overcharges, provides policy‑cited explanations.
  • Stack: PaddleOCR/Tesseract, LayoutLM/Donut, rule/LLM hybrid matching, LangChain RAG, Streamlit, pytest, Docker.
  • Outcome focus: High‑precision auto‑matches, exception review time reduction, explainability with citations.

• MMM + Uplift Budget Optimizer (Marketing Science)

  • What it does: Bayesian MMM + uplift modeling for channel allocation; generates marketer‑friendly, evidence‑backed rationales.
  • Stack: PyMC/Orbit, LightGBM, Pandas, Streamlit, LangChain (brand/policy RAG), MLflow.
  • Outcome focus: Out‑of‑time ROI error, Qini/AUC for uplift, scenario analysis with assumptions documented.

• Telecom Churn ML Pipeline

  • What it does: Config‑driven pipeline with single CLI for FE → prep → CV/tuning → eval → best‑pick → batch scoring.
  • Stack: Python, scikit‑learn, Ensembling, Hyperparam tuning, MLflow/DVC, CI on GitHub Actions.
  • Outcome focus: Reproducible artifacts, thresholded risk outputs, campaign deciles for actionability.

• Predictive Maintenance (Tabular ML)

  • What it does: Random Forest/XGBoost for machine failure prediction with feature selection and SMOTE.
  • Outcome focus: Deployment‑ready artifacts and monitoring hooks; >90% accuracy on benchmark dataset.

Want links? I’m happy to make these repos public or provide private demos on request.

• BI Tools: Qlik, Tableau, Power BI, Looker
• Languages & Data Wrangling: Python (NumPy, Pandas, Matplotlib, Seaborn), R, SQL
• ML & AI Frameworks: scikit-learn, TensorFlow, PyTorch, Hugging Face Transformers, LangChain, LlamaIndex, spaCy
• MLOps & Deployment: AWS (SageMaker, EC2, S3), Docker, Git/GitHub, MLflow, Weights & Biases, FastAPI, Streamlit
• Soft Skills: Problem Solving, Communication, Collaboration, Leadership, Stakeholder Alignment, Documentation, Delivery Discipline

• Manito Mix — Digital Marketing Analyst (2023–2024)

  • SEO audits across 20+ sites → +35% organic traffic in 3 months
  • A/B tests + funnel diagnostics → +40% conversions
  • $150K+ ads → 2.4× ROAS, lower CPA; automated reporting → −60% time

• Kissflow — Digital Marketing Associate (2022–2023)

  • 15+ SEO audits → +50% non‑branded traffic; $200K+ PPC → 3× ROAS
  • 30+ backlink campaigns → +12 DA; CRO improvements → +37% lead conversion

• Procter & Gamble — Supply Network Ops Intern (2021)

  • SKU‑level forecasting → −12% excess inventory, +7% service levels
  • Ordering strategy + transport model → +8% CFR, better DOC alignment

• Northeastern University — MPS Analytics, Applied Machine Intelligence (2025–2027) • CGPA 4.0
• NUST — BS Accounting & Finance, Minor in Economics (2018–2022)

• Google Advanced Data Analytics Professional Certificate (2025)
• Global Leaders Project Intern – Youth Thinkers’ Society & U.S. Embassy, Kathmandu, Nepal (2017)
• Certificate of Achievement – International Economic Summit, Spokane, WA, USA (2016)
• Kennedy-Lugar Youth Exchange Ambassador – Senior Year at West Valley High School, Spokane, WA (2015–2016)

• Email: [email protected]
• LinkedIn: linkedin.com/in/muhammad-o
• GitHub: github.com/muhammad-o98