Thanseefuddeen

Data Scientist|ML Engineer|LLM & GenAI Specialist

Building production-grade AI systems that bridge the gap between research and reality — fine-tuned LLMs, computer vision pipelines, and MLOps at scale. Currently a Data Scientist at Orion Innovation.

5+ Years ExperienceTurning Research into Real ProductsDriven by Impact, Not Just Code
Thanseefuddeen — Data Scientist & ML Engineer

01 // Expertise

Applied Machine Learning & AI

I'm a Data Scientist with 5+ years across machine learning, deep learning, and full-stack engineering. I've shipped LLM fine-tuning, multi-agent systems, computer-vision products, and NL2SQL analytics for teams in the US, Singapore, and the UAE — owning problems end to end, from research to production.

Generative AI / LLMs

Fine-tuning (Qwen 2.5, Llama 3, Unsloth, LoRA, QLoRA), RAG, and multi-agent systems with source-cited retrieval.

Computer Vision

Face recognition & liveness (InsightFace, ONNX), YOLOX detection, and image classification on low-end hardware.

MLOps & Deployment

Reproducible pipelines with MLflow & ZenML, containerized with Docker, deployed on AWS, GCP, and Azure.

Backend & Data

FastAPI, Django & Node.js services; Neo4j & MongoDB; parallel processing with Celery, Redis, and RabbitMQ.

02 // Skills

Technical Stack

Languages
PythonSQLJavaScriptTypeScriptR
Core ML / DL
TensorFlowPyTorchScikit-LearnKerasOpenCVYOLOXRoboflowPandasNumPy
LLMs & GenAI
TransformersLangChainUnslothLoRAQLoRARAGOpen ModelsWhisperDreamboothNL2SQLLiveKitMCPFastMCPMicrosoft Agent FrameworkAzure OpenAIAWS Bedrock
MLOps & Tooling
MLflowZenMLBentoMLKedroDockerJenkinsGitHub ActionsGitJupyterClaude Code
Observability & Eval
DeepEvalLangfuseTensorBoardOpenTelemetrySentry SDK
Backend
FastAPIDjangoFlaskNode.jsNestJSCeleryRedisRabbitMQ
Frontend & Web
Next.jsReactGSAPLenisStripeTailwind CSS
Mobile
FlutterSwift
Cloud · AWS
SageMakerECSLambdaS3RDSCloudWatchElastic BeanstalkBedrockOpenSearch
Cloud · Azure
FunctionsAI FoundryContainer AppsVMData LakeMicrosoft FabricKey VaultService BusAI SearchCosmos DB
Cloud · GCP
Vertex AIBigQueryCloud StorageVM Instances
Databases
Neo4jMongoDBPostgreSQLMySQLQdrantSingleStorePineCone
Data Viz
TableauPower BIMatplotlibSeaborn

03 // Experience

Work History

Mar 2025 — Present

Data Scientist

Orion Innovation · Hybrid · Kochi

  • →Finetuned a Qwen 2.5 8B-parameter model with Unsloth on 15k+ custom training samples to encode organizational policies, after evaluating multiple large language model architectures.
  • →Built an intelligent customer feedback classification system for American Airlines using a Neo4j graph database and agent-based inference, automating categorization of user queries into actionable business segments.
  • →Engineered a multi-agent recommendation platform for IDEX Corp that suggests optimal pump solutions with source citations from videos and documentation, integrated with ML-based match-score prediction for catalog ranking.
Aug 2023 — Jan 2025

Machine Learning Engineer

Fuelgrowth · Remote · Singapore

  • →Designed Fugo Analytics, an AI-powered NL2SQL tool, driving a 30% increase in reporting accuracy across platforms like Shopify and Meta.
  • →Deployed a Random Forest sales-forecasting model within AI pipelines, achieving 92% accuracy and driving strategic insights.
  • →Adapted image and text generation models using Dreambooth and Llama 3 for personalized marketing campaigns, and automated customer data into tailored creatives via a GCP pipeline.
Feb 2023 — Aug 2023

Machine Learning Engineer (Research Assistant)

Zayed University · Remote · Dubai

  • →Engineered an AI platform integrating machine learning, deep learning, and computer vision, using LLMs to boost query-matching accuracy by 25%.
  • →Launched realtime lip-synchronization connecting agent voice output with animated avatar visuals for an immersive, human-like learning experience.
  • →Configured a YOLOX-based emotion-recognition model on Roboflow datasets for real-time engagement analytics, and accelerated workflow efficiency by 80% with MLflow and ZenML.
Jan 2022 — Dec 2022

Software Engineer

Eclidse Technologies · Calicut · Kerala

  • →Architected a smart review filtering system using NLP algorithms, reducing manual review time by 25%.
  • →Accelerated data processing by 40% through Python parallel processing with Celery, Redis, and RabbitMQ.
  • →Enhanced website performance by configuring auto-scaling and load balancing on EC2, reducing downtime by 50% and customer complaints by 10%.
Jan 2021 — Jan 2022

Jr. Software Engineer

Saasvaap Inc · Remote · Kerala

  • →Collaborated with designers, developers, and clients in remote settings, ensuring seamless delivery of web applications.
  • →Integrated RPA solutions using Jiffy.ai, improving customer satisfaction by 10%.
  • →Resolved ERP system performance issues, supporting a 60% increase in concurrent users.
Jun 2019 — Nov 2019

Network Engineer (L1)

IBM India Pvt. Ltd. · On-site · Bengaluru

  • →Operated within the NOC team, configuring Cisco routers and switches remotely over PuTTY in coordination with on-site field engineers across client locations.
  • →Monitored client network health, uptime, and incidents using SolarWinds, tracking issues and downtime to maintain service continuity.
  • →Maintained network inventory and issue records to support faster L1 triage and escalation.

04 // Selected Projects

Featured Projects

ZenAI
Voice AI · LLM1000 → 300ms

ZenAI

Configurable realtime voice agent for inbound & outbound customer calls, integrated with Telnyx SIP and deployed on AWS.

FastAPILiveKitNext.jsLLMSTT/TTSMongoDB
Hodoor AI
Computer Vision−40% latency

Hodoor AI

Scalable face-recognition attendance system with liveness detection, optimized for low-end devices via ONNX conversion.

OpenCVInsightFaceONNXFastAPI
Deep Learning96% acc

Agroctor

Deep-learning pest & leaf-disease classifier paired with an AI agent that recommends treatments — reducing crop damage.

TensorFlowRoboflowWhisperLLM
Fitzy
Deep Learning

Fitzy

Apparel recommendation using deep learning and computer vision.

Deep LearningComputer Vision
AceResume
AI · LLM

AceResume

AI-powered resume optimizer built on LLMs and NLP.

AILLMNLP
Gemstone Price Prediction
Machine Learning

Gemstone Price Prediction

End-to-end ML application with an MLOps deployment pipeline.

MLRegressionMLOps
View All Projects on GitHub

05 // Contact

Open to Data Science & ML roles

thanseefpp@gmail.com

Location

Calicut, Kerala, India

GitHubLinkedInTwitterLeetCode