I am Anmol Tripathi, a Quality Data Scientist working across data science, machine learning, applied AI, analytics engineering, and business-facing automation. My work spans the full path from data preparation and experimentation to model evaluation, reproducible inference, CI validation, and interactive deployment.
This portfolio documents a deliberate technical progression: applied analytics → classical machine learning → deep neural networks → sequence models → Transformers → deployed AI applications.
Current focus: reliable NLP, retrieval, multimodal AI, model benchmarking, and production-oriented ML workflows—built with transparent evaluation and responsible-use documentation.
|
10 completed and deployed projects spanning summarization, translation, semantic retrieval, reranking, long-document QA, instruction tuning, visual question answering, Vision Transformers, CLIP, and RAG.
|
End-to-end encoder–decoder applications including schema-aware Text-to-SQL, streaming-style speech recognition, and data-to-text executive reporting.
|
|
A portfolio of analytical and predictive workflows demonstrating problem framing, preprocessing, EDA, statistical reasoning, feature engineering, model comparison, and business communication.
|
Computer-vision projects organized around convolutional architectures, data pipelines, model training, evaluation, and visual error analysis.
|
Five representative applications selected to demonstrate breadth across structured generation, retrieval-augmented AI, computer vision, speech recognition, and business-facing analytics.
| Project | What it demonstrates | Explore |
|---|---|---|
| Schema-Aware Text-to-SQL | Fine-tunes CodeT5+ 770M with LoRA to generate SQLite from natural-language questions and database schemas, then validates and safely executes approved read-only queries. | Source code · Live application |
| AI Portfolio RAG Assistant | Searches public portfolio evidence with MiniLM embeddings and hybrid retrieval, returning grounded answers with source citations, relevance evidence, and latency details through a Next.js application. | Source code · Live application |
| Vision Transformer Browser Classifier | Compares a DeiT-tiny Vision Transformer with ResNet-18, validates PyTorch-to-ONNX parity, visualizes attention rollout, and performs private WebGPU/WASM inference in the browser. | Source code · Live application |
| Streaming Speech Recognition with Whisper | Transcribes microphone or uploaded audio using a Whisper encoder-decoder workflow with chunking, timestamps, language detection, and robustness-oriented evaluation. | Source code · Live application |
| Data-to-Text Executive Report Generator | Converts structured KPI tables into source-grounded executive narratives while checking numerical claims, surfacing source-cell evidence, and blocking unsupported statements. | Source code · Live application |
| Capability | Evidence in this portfolio |
|---|---|
| Applied Data Science | EDA, statistical analysis, feature engineering, predictive modeling, business-facing visualization |
| Classical Machine Learning | Classification, regression, clustering, ensemble modeling, benchmarking, calibration, error analysis |
| Deep Learning | ANN, CNN, RNN, LSTM, bidirectional LSTM, encoder–decoder, and Transformer architectures |
| NLP & Retrieval | Summarization, translation, QA, semantic search, dense retrieval, cross-encoder reranking, RAG |
| Computer Vision & Multimodal AI | CNNs, Vision Transformers, visual question answering, CLIP image–text retrieval |
| ML Engineering | Reproducible training, modular inference, GPU/BF16 workflows, ONNX conversion, testing, CI/CD, deployment |
| Analytics Engineering | Python/SQL automation, structured reporting, Power BI, and decision-oriented communication |
Applied analytics & statistics
↓
Classical machine learning
↓
Artificial neural networks & computer vision
↓
RNN, LSTM & bidirectional LSTM systems
↓
Encoder–decoder architectures
↓
Transformers, multimodal AI & retrieval-augmented generation
↓
Evaluation, testing, CI/CD & deployment
Languages & analysis
Machine learning & deep learning
Delivery & visualization
My professional work combines quality data, operational context, analytics, automation, and machine learning. I build workflows that help teams move from fragmented data to repeatable analysis and decision support—including data preparation, classification, model validation, dashboarding, and automated reporting.
Public repositories contain portfolio work and non-confidential demonstrations only. Proprietary data, internal systems, and company methodology are intentionally excluded.
I apply data science, machine learning, analytics engineering, and automation to non-confidential quality and operational problems. My work includes developing automated data pipelines, predictive classification workflows, SQL-based reporting, Power BI dashboards, data-source integration, and reusable analytical processes that support quality assurance and operational decision-making.
Technical focus: Python · SQL · Power BI · Machine Learning · Statistical Analysis · Predictive Modeling · Data Automation · Model Validation · Reporting Workflows
Company datasets, internal systems, proprietary source code, and confidential methodologies are not published on this profile. The public repositories demonstrate related technical capabilities using independent or publicly permitted data.
Applied deep-learning architectures—including LSTM, bidirectional LSTM, CNN, and autoencoder-based approaches—to healthcare research involving labor-duration prediction. The work included data preparation, model experimentation, performance evaluation, and communication of findings within an interdisciplinary research environment.
Technical focus: Python · TensorFlow/Keras · LSTM · Bidirectional LSTM · CNN · Autoencoders · Healthcare Analytics · Model Evaluation
- Frame the decision — define the real problem, target, constraints, and meaningful evaluation criteria.
- Establish evidence — validate data, build baselines, compare candidates, and inspect failure modes.
- Engineer for reuse — separate training and inference, preserve metadata, test artifacts, and document assumptions.
- Deliver responsibly — expose limitations, validate deployments, and communicate results for technical and business audiences.
|
University of Arizona Master of Science in Data Science GPA: 3.889 / 4.000 |
Texas McCombs School of Business Post Graduate Program in Data Science and Business Analytics Overall grade: 4.00 / 4.00 |
University of Arizona — Curriculum and academic performance
Academic summary: 30 graduate units · 8 A grades · 1 B grade · 1 Pass
| Term | Course | Units | Grade |
|---|---|---|---|
| Fall 2022 | INFO 520 — Ethical Issues in Information | 3 | A |
| Fall 2022 | INFO 521 — Intro to Machine Learning | 3 | B |
| Fall 2022 | INFO 523 — Data Mining/Discovery | 3 | A |
| Spring 2023 | INFO 507 — Information Research Methods | 3 | A |
| Spring 2023 | INFO 526 — Data Analysis and Visualization | 3 | A |
| Spring 2023 | INFO 550 — Artificial Intelligence | 3 | A |
| Fall 2023 | INFO 555 — Applied NLP | 3 | A |
| Fall 2023 | INFO 557 — Neural Networks | 3 | A |
| Fall 2023 | INFO 579 — SQL/NoSQL Databases | 3 | A |
| Fall 2023 | NURS 599 — Independent Study | 3 | Pass |
Texas McCombs — Data Science and Business Analytics curriculum
| Learning area | Coursework |
|---|---|
| Programming and data | Python for Data Science · SQL |
| Statistics and analytics | Statistical Methods for Decision Making · Advanced Statistics · Data Mining |
| Predictive modeling | Predictive Modeling · Machine Learning · Time Series Forecasting |
| Business applications | Marketing and Retail Analytics · Finance and Risk Analytics |
| Visualization | Data Visualization using Tableau |
| Applied integration | Capstone Project — Post Graduate Program in Data Science and Business Analytics |
I continue to expand this portfolio toward production-oriented AI and ML engineering, with current emphasis on:
- retrieval-augmented generation, semantic retrieval, reranking, and evidence-grounded AI;
- instruction tuning, LoRA, and parameter-efficient model adaptation;
- multimodal systems, ONNX conversion, and browser-based inference;
- model evaluation, testing, CI/CD, cloud deployment, monitoring, and model versioning.
- [Official certification title] — [Issuing organization] · [Credential link]
- [Official certification title] — [Issuing organization] · [Credential link]
- [Official certification title] — [Issuing organization] · [Credential link]
| Path | Best starting point |
|---|---|
| AI / ML Engineering | Transformer projects → testing, CI, ONNX, browser inference, deployed applications |
| NLP / Generative AI | Encoder–decoder projects → Text-to-SQL, speech recognition, data-to-text systems |
| Data Science / Analytics | Applied DS & ML portfolio → analysis, modeling, evaluation, business communication |
| Computer Vision | CNN projects and Vision Transformer work |
| Sequence Modeling | RNN → LSTM → BiLSTM |
Interested in data science, ML engineering, or applied AI?
Connect with me on LinkedIn