MANTHAN C — Senior AI/ML Engineer

Senior AI/ML Engineer - Cyber Infrastructure (CIS) Pvt. Ltd. (Enterprise RAG Knowledge Assistant)

Duration: Feb 2021-Present

Summary:

• Built an enterprise Retrieval-Augmented Generation (RAG) platform that answers questions from company knowledge using hybrid retrieval combining vector search and graph traversal
• The platform leverages Neo4j to model relationships between documents and entities for enhanced context and uses multiple open-source LLMs with multi-stage reranking to improve answer quality

Responsibilities:

• Designed and implemented hybrid retrieval system using vector databases and graph databases.
• Developed routing mechanisms across 50+ open-source LLMs and multi-stage reranking with cross-encoders.
• Implemented hallucination reduction techniques by enforcing structured output.
• Built authentication, monitoring, and evaluation suites to benchmark LLM quality.

Technologies: Python, FastAPI, LangChain, Neo4j, Hugging Face, FAISS, Runpod

Senior AI/ML Engineer - Cyber Infrastructure (CIS) Pvt. Ltd. (PersonaGrid: AI Sales Enablement Platform)

Duration: Feb 2021-Present

Summary:

• Developed an AI platform for sales call simulation and pre-call planning that supports multi-tenant data isolation and structured AI output
• The platform enables sales teams to prepare effectively by simulating calls and reasoning through multi-step AI pipelines

Responsibilities:

• Designed multi-step reasoning pipelines using LangChain chaining.
• Implemented multi-tenant vector database architecture to ensure data separation per organization.
• Enforced JSON schema to maintain structured and compliant AI outputs.
• Managed background task execution with Celery and Redis and automated organizational setup through CI/CD pipelines.

Technologies: Python, FastAPI, LangChain, OpenAI, Celery, Redis, Next.js, AWS

Senior AI/ML Engineer - Cyber Infrastructure (CIS) Pvt. Ltd. (AI-Powered Data Protection Compliance Portal)

Duration: Feb 2021-Present

Summary:

• Built a GDPR compliance portal that automates data protection queries and risk audits using AI
• The portal assists compliance teams by providing AI-assisted PII redaction, severity scoring, and automated responses to reduce manual workload

Responsibilities:

• Developed AI pipelines for PII redaction, severity scoring, and automated query responses using Claude 3.
• Implemented audit scoring and visual reporting features for compliance teams.
• Integrated authentication with Auth0 and deployed on AWS infrastructure.

Technologies: Python, FastAPI, LangChain, ChromaDB, Claude 3, Auth0, AWS

Senior AI/ML Engineer - Cyber Infrastructure (CIS) Pvt. Ltd. (Semantic Text & Visual Search Engine)

Duration: Feb 2021-Present

Summary:

• Developed a semantic search engine supporting both text and image queries using embedding-based retrieval
• The system enables multimodal search capabilities and scalable model deployment for efficient and accurate search results

Responsibilities:

• Built embedding-based retrieval system with FAISS indexing.
• Implemented multimodal search supporting text and image queries.
• Developed image similarity pipelines with experiment tracking for model evaluation.
• Deployed and managed models on AWS SageMaker with monitoring and scalable inference.

Technologies: Python, TensorFlow, Hugging Face, FAISS, AWS SageMaker, S3, DynamoDB

Senior AI/ML Engineer - Cyber Infrastructure (CIS) Pvt. Ltd. (Horse Race Outcome Prediction System)

Duration: Feb 2021-Present

Summary: Built a machine learning system to predict horse race outcomes including final position and return on investment (ROI) using ensemble models trained on large historical datasets.

Responsibilities:

• Developed stacked ensemble models using LightGBM and CatBoost.
• Designed feature engineering and validation pipelines to ensure data quality and reproducibility.
• Conducted repeated experiments and model tuning to improve prediction accuracy.

Technologies: Python, scikit-learn, LightGBM, CatBoost, Pandas, NumPy

Senior AI/ML Engineer - Cyber Infrastructure (CIS) Pvt. Ltd. (OCT Medical Image Segmentation)

Duration: Feb 2021-Present

Summary:

• Developed a desktop application for segmenting optical coherence tomography (OCT) medical eye scan images using deep learning
• The application assists medical professionals by providing accurate segmentation and measurement extraction

Responsibilities:

• Developed a custom U-Net model for mask prediction on OCT images.
• Used OpenCV for image masking and measurement extraction.
• Applied transfer learning to improve classification accuracy.
• Wrapped the model in a PyQt interface for desktop application usability.

Technologies: Python, TensorFlow, OpenCV, U-Net, Roboflow, PyQt