Aniket Poojari

AI Researcher

United States1 yr 8 mos experience

AI EnabledAI ML Practitioner

Key Highlights

3× faster diffusion model training using DeepSpeed.
20% higher forecasting accuracy through automated ML workflows.
On-premise RAG chatbot reducing support resolution time.

Stackforce AI infers this person is a Machine Learning Engineer with expertise in Fintech and Healthcare sectors.

Contact

aniketp2009@gmail.com LinkedIn

Skills

Core Skills

Machine LearningData ScienceData EngineeringSoftware Development

Other Skills

AJAXAWS SageMakerAirflowAlgorithmsAmazon S3AndroidApache AirflowApache SparkArduinoC++Cascading Style Sheets (CSS)CloudinaryComputer VisionContinuous Integration and Continuous Delivery (CI/CD)Contrastive Language-Image Pre-training

About

I’m a Machine Learning Engineer with 3+ years of experience building production-grade AI systems across fintech, EdTech, and research environments focused on performance, reliability, and measurable impact. I specialize in taking AI from prototype to production, distributed training, scalable LLM/RAG systems, and ML infrastructure that reduces latency, cost, and manual work. I’ve delivered systems used across trading firms, SaaS platforms, and academic ML programs. What I’m good at • Real-time AI pipelines (LLMs, RAG, agentic AI, generative models) • Distributed model training (PyTorch, DeepSpeed, AWS/GCP) • Secure on-premise & cloud ML systems (Docker, Kubernetes, LangGraph) • Full-stack ML deployment (FastAPI, Express, React, CI/CD) • Turning research ideas into production systems with business impact Career highlights • 3× faster diffusion model training via a distributed DeepSpeed pipeline on AWS SageMaker. • 20% higher forecasting accuracy + 15% fewer allocation errors through automated ML workflows. • On-premise RAG chatbot reducing support resolution from 3 days → 1 day. • Agentic research assistant (LangGraph) with 99.9% test accuracy, cutting workflows from hours to minutes. • ML systems for finance/computer vision, reducing processing time by 30–45%. What drives me If a process is slow, repetitive, or expensive, I make it fast, scalable, and automated. Call to action If you're building AI systems where scalability, latency, or reliability matter, I’d love to connect.

Experience

1 yr 8 mos

Total Experience

1 yr 8 mos

Average Tenure

Current Experience

Kritova

Machine Learning Engineer

Mar 2025 – Present · 1 yr 3 mos · United States · Remote

Futuretech foundation

Software Engineer (Volunteer)

Feb 2025 – Present · 1 yr 4 mos · United States · Remote

Built a Python pipeline to clean and transform brainwave recordings into features and trained machine learning models to predict which movement a person is imagining (left hand, right hand, foot, tongue) from EEG signals.
Developed reusable Python tools to load, standardize, and label brainwave data from OpenBCI headsets and public datasets, creating training/evaluation datasets and iterating on baseline classifiers to improve imagined-movement detection accuracy.

PythonMachine LearningData CleaningFeature EngineeringData Science

Oregon state university

Graduate Teaching Assistant

Mar 2023 – Dec 2024 · 1 yr 9 mos · Corvallis, Oregon, United States · On-site

Graduate Teaching Assistant for Applied Machine Learning, Cloud Application Development, Mobile Software Development, Open Source Software, and Web Development

Dolat capital

Software Developer

Aug 2020 – Apr 2022 · 1 yr 8 mos · Mumbai, Maharashtra, India · On-site

Built production ML pipelines for block-trade allocation and forecasting, improving forecast accuracy by 20% and reducing allocation errors by 15%.
Automated data infrastructure with PostgreSQL + Docker, boosting data retrieval speed by 75% and cutting system downtime by 30%.
Optimized large-scale Apache Spark workflows used in live/sim trading, increasing testing consistency by 50% and reducing data discrepancies by 35%.
Automated trade-cleansing workflows (Random Forest, XGBoost), reducing manual workload by 30% and improving model reliability by 25%.
Built Python/Bash automation for trade metrics (Sharpe, PnL parsing), accelerating model evaluation cycles by 15×.
Created interactive analytics dashboards to give traders faster insights into stock behavior and model performance.
Maintained high-reliability C++ trading components for production environments, ensuring low-latency performance and strict security compliance.
Replaced manual data sourcing with automated parameter retrieval from distributed servers, accelerating model-vs-sim comparisons.
Collaborated with quantitative analysts to develop custom ML tools, improving analytical workflows across the trading desk.

Machine LearningPostgreSQLDockerApache SparkData Engineering

Dhfl

Artificial Intelligence Intern

Jun 2019 – Jul 2019 · 1 mo · Mumbai, Maharashtra, India · On-site

Automated vehicle damage detection for insurance claims by retraining Mask R-CNN, improving detection accuracy by 18% and reducing manual claim entry by 45%.
Integrated ML workflows into a Django-based system, cutting claim processing time by 35% and streamlining end-to-end approvals.
Managed large-scale preprocessing of customer claim images, ensuring high-quality training data for reliable ML model performance.
Delivered a production-ready solution combining TensorFlow, OpenCV, Django, and web technologies, enabling seamless image uploads and faster claim approvals.

Machine LearningImage ProcessingDjangoSoftware Development