Services
AI Chatbots & Virtual Assistants
I build production-grade chatbots with RAG pipelines, fine-tuning, and domain-specific LLMs for precise, context-aware responses. Integrations include help desks, CRMs, and internal tools—shipped with auth, rate limits, analytics, and fallbacks for reliability.
Generative AI Modeling & Local AI
I deliver custom and composable gen-AI features (images, text, audio, code). Where privacy/costs matter, I deploy local or VPC-isolated models to keep data in-house. Credit metering, caching, and observability ensure predictable performance and spend.
Natural Language Processing (NLP)
From embeddings to classic ML, I implement sentiment analysis, summarization, topic modeling, and recommendation systems. Data pipelines, evaluation harnesses, and A/B testing are standard so models stay accurate and useful over time.
Computer Vision
I build CV workloads like detection, OCR, and segmentation using YOLO/R-CNN/OpenCV. Deployed as scalable APIs with proper batching, GPU utilization, and latency budgets—plus guardrails and QA datasets for consistent real-world accuracy.
Predictive Modeling & Time Series
I implement forecasting, anomaly detection, and causal features for KPIs and operations. Expect clean feature stores, backtests, and drift monitoring so predictions remain trustworthy as data and seasonality shift.
Reinforcement Learning
For sequential decision problems, I prototype and productionize RL agents with safe reward design, simulation loops, and offline evaluation. Telemetry, replay buffers, and guardrails keep learning stable and auditable.
Experiences
Feb 2023 - Apr 2025
Feb 2023 - Apr 2025
AI/ML Engineer
@DataNova Analytics
Designed and deployed ML solutions across finance, healthcare, and retail.
Built custom NLP pipelines for sentiment analysis and entity recognition.
Developed time-series forecasting models (ARIMA, LSTM, Prophet).
Used TensorFlow, PyTorch, and Scikit-learn for deep learning projects.
Led model deployments using AWS EC2 and Docker.
Collaborated cross-functionally to deliver data-driven products.
Dec 2021 - Jan 2023
Dec 2021 - Jan 2023
Machine Learning Research Associate
@VisionTech Labs
Built object detection systems using YOLOv4 and R-CNN.
Experimented with GANs for synthetic image generation.
Published internal docs on model optimization and transfer learning.
Worked with PhDs and research engineers on deep learning experiments.
Assisted in deploying models on embedded and edge AI hardware.
Oct 2020 - Nov 2021
Oct 2020 - Nov 2021
Junior Data Scientist
@ByteShift Solutions
Cleaned and structured large datasets using Pandas and NumPy.
Built regression and classification models for internal projects.
Assisted senior data scientists with feature engineering and evaluation.
Worked on early NLP tasks like text classification.
Created data visualization dashboards with Matplotlib and Seaborn.
Skills
Programming Languages
Python
TypeScript
JavaScript
AI/ML Frameworks
PyTorch
TensorFlow
scikit-learn
Keras
Hugging Face
NLP & Computer Vision
spaCy
OpenAI
OpenCV
Data Science
NumPy
Pandas
Jupyter
SciPy
Plotly
Databases
PostgreSQL
MongoDB
MySQL
Redis
SQLite
DevOps & MLOps
Docker
Kubernetes
Terraform
MLflow
Apache Airflow
Cloud & Hosting
Google Cloud
Vercel
Netlify
APIs & Developer Tools
Node.js
Express
Postman
Git
GitHub
Recent Works
Voice-first mock interview system that plans topics from candidate intake, runs a live VAPI session, and outputs rubric-scored feedback with a study plan (Gemni orchestration).
Gemni
VAPI
SaaS for object removal, recolor, background change, generative fill, and restore. Deterministic params, NSFW guardrails, and credit billing with Stripe; Cloudinary Generative under the hood.
Cloudinary
Redis
BullMQ
RAG-first chatbot that answers policy & doctor-finder queries with citations, tool calling for slot peek & prefilled booking links, and strict non-diagnostic guardrails.
RAG
OpenAI
Qdrant
CV-first pipeline with OpenCV preprocessing, engine routing (Tesseract/Paddle), confidence scoring, regex/dictionary post-correction, and optional LLM span cleanup.
OpenCV
Tesseract
PaddleOCR
LLM
Uploads & parses PDFs, indexes chunks with page anchors, and answers with streaming responses and page-level citations; table extraction and optional OCR for scans.
RAG
OpenAI
Qdrant
SSE
Get in Touch
I'm available for freelancing.