Available to hire
Hi, I’m Sahithi M. I am an AI Engineer with over 3 years of experience designing and deploying production-grade GenAI and machine learning solutions using Python, LLMs, and RAG architectures. I enjoy building scalable, cloud-ready AI systems with FastAPI, PyTorch, and LlamaIndex, and I love integrating multimodal reasoning across text, image, and 3D data to deliver practical business outcomes.
I thrive in fast-paced, collaborative environments, designing multi-agent workflows, structured prompts, and evaluator mechanisms to guide reasoning and reduce hallucinations. I’m passionate about turning complex problems into reliable AI-powered solutions that drive business impact.
Language
English
Advanced
Work Experience
AI Engineer at Innovision LLC
May 1, 2025 - PresentDeveloped ComFit Copilot, a production-grade multimodal GenAI system enabling intelligent retrieval across text, images, and 3D models. Architected multi-agent workflows and multi-step/pipeline strategies for adaptive reasoning on complex queries. Built hybrid RAG pipelines using DuckDB, ChromaDB, and Qdrant for offline/online retrieval with low-latency vector search. Implemented Retrieval-Augmented Correction and confidence scoring to validate claims and reduce hallucinations. Evaluated multiple embedding models and achieved high retrieval accuracy. Designed MCP-style tool schemas and context-sharing logic for consistent agent communication, and integrated a dual-mode image retrieval pipeline with CLIP-based search and 3D model visualization via Sketchfab. Optimized platform with authentication, secure storage, UI reasoning-step viewer, environment configuration, and caching for a robust AI experience.
Data Scientist at Protominds Software Solutions
May 1, 2022 - August 1, 2024Built an end-to-end ML-driven credit loss prediction system using Vintage, Snapshot, and WARM methodologies to estimate lifetime charge-off rates and losses. Engineered financial features (Default UPB, Delinquent Accrued Interest, Actual Loss) from large loan-performance data and trained models (Bayesian HMM, XGBoost, Random Forest, SVM, Logistic & Linear Regression) to predict charge-off probability and loss severity. Applied statistical and ML methods to identify risk segments, and automated OLS/Bayesian regression-based loss forecasting to project future trends. Produced dashboards showcasing historical, predicted, and adjusted loss metrics to guide risk strategy.
Data Science (Vintage, Snapshot, and WARM Frameworks) at Protominds Software Solutions
May 1, 2022 - August 1, 2024Built an end-to-end ML-driven credit loss prediction system using Vintage, Snapshot, and WARM (Weighted Average Remaining Maturity) methodologies to estimate lifetime charge-off rates and credit losses. Engineered key financial features such as Default UPB, Delinquent Accrued Interest, and Actual Loss from large-scale loan performance data using advanced statistical formulas and time-series aggregation. Performed extensive data cleaning and transformation with Pandas and NumPy, handling missing values, normalization, and temporal joins. Trained models (Bayesian HMM, XGBoost, Random Forest, SVM, Logistic & Linear Regression) to predict charge-off probability and loss severity across cohorts and product segments. Applied statistical/machine learning methods (Decision Trees, Naive Bayes, PCA, Clustering) to identify patterns, risk segments, and drivers of default behavior. Automated OLS and Bayesian regression-based extrapolation to forecast future loss trends, replacing manual Excel TREN
Education
Master of Science in Computer Science at University of Cincinnati
August 1, 2024 - December 1, 2025Master of Science in Computer Science at University of Cincinnati
August 1, 2024 - December 1, 2025Qualifications
Industry Experience
Software & Internet, Professional Services, Media & Entertainment, Financial Services
Hire a AI Engineer
We have the best ai engineer experts on Twine. Hire a ai engineer in Dayton today.