Manideep Sriperambudhuru

Applied AI / ML Engineer

Hyderabad, India manideepsp.16@gmail.com LinkedIn
GenAI EngineerLLM SystemsML PipelinesData Engineering

Applied AI / ML Engineer with nearly three years of experience spanning data engineering, applied machine learning, and GenAI systems. Strong background in building reliable data pipelines and analytics foundations, followed by hands-on ownership of ML- and LLM-driven systems including RAG pipelines, agentic chatbots, and hyper-personalized learning workflows. Experienced in taking internal proof-of-concepts from data ingestion and modeling through experimentation, optimization, deployment, and platform integration.

Skills & Expertise

GenAI / LLM Systems

9 skills
LoRA and PEFT (Parameter-Efficient Fine-Tuning)
Agentic AI systems with stateful, multi-step reasoning
LangChain, LangGraph, and LangSmith
RAGAS evaluation framework
Relevance, groundedness, and faithfulness checks
LLM latency and cost optimization
Prompt design, refinement, and token reduction
Prompt evaluation using cosine similarity and BLEU score
Retrieval-Augmented Generation (RAG)

Core Competencies

GenAI / LLM Systems

LoRA and PEFT fine-tuning techniquesAgentic conversational systems (LangGraph)RAGAS evaluation frameworkRelevance and faithfulness evaluationBLEU and similarity-based evaluation techniquesLLM latency and cost optimization

Machine Learning

Handling temporal and sequential data patternsCross-validation, metric selection, and tradeoff analysisModel evaluation and experimentationFeature engineering from structured and unstructured dataClassical ML (regression, clustering)

Engineering & Platform

Internal CI/CD exposure and deployment workflowsDocker-based containerizationAzure ecosystem (ADF, Databricks, Storage, Managed Identities)REST-based service integrationPython and SQL

Data Engineering & Analytics Foundations

BI reporting and stakeholder-facing analyticsData quality checks, logging, and error handlingSQL query optimization and stored proceduresData modeling for analytics and ML workloadsMedallion Architecture (Bronze / Silver / Gold layers)ETL pipeline design and orchestration

Experience

Cognine TechnologiesApplied AI / ML Engineer
April 2023 – December 2025Hyderabad, India

Role Overview

My role at Cognine Technologies evolved from a strong data engineering foundation into applied ML and GenAI engineering, delivering multiple production-grade systems.

Primary Goals

  • Design and deliver AI/ML and GenAI systems with measurable impact
  • Build scalable data foundations for analytics and AI
  • Optimize latency, cost, safety, and reliability of LLM workflows
  • Collaborate with cross-functional teams for production readiness

Core Responsibilities

  • Client Projects: Delivered assigned development components across data engineering, ML, and GenAI as part of multi-disciplinary teams.
  • Internal AI Projects: Owned the full lifecycle — requirements, architecture, backend/frontend development, CI/CD, DevOps, and production deployment.
  • System architecture, performance tuning, and production hardening
  • Evaluation, monitoring, and safety of AI systems
  • Integration of AI/ML systems with backend services and apps

Overall Experience Summary

Across all projects, I contributed to data engineering, applied ML, and GenAI, with increasing emphasis on AI-first problem solving and end-to-end delivery.

AI Tutoring Platform — LLM Reports, RAG & Agentic Chatbot

Nov 2025

Focus Area: GenAI → ML → Data

Objective

Build a safe, scalable, and personalized AI tutoring platform using LLMs, RAG, and agentic architectures for under-18 students.

LLM-Based Report Generation

  • Developed AI-generated Basic and Premium study reports
  • Broke monolithic prompts into section-wise generation
  • Executed parallel LLM calls using multithreading
  • Applied prompt and token optimization techniques

Results

  • Reduced report generation from ~4.5 minutes → ~45 seconds
  • Significantly lowered token consumption and operating cost

Agentic Conversational AI (Chatbot)

  • Built an agentic chatbot using LangGraph with multi-agent collaboration
  • Added persona profiling and mood tracking
  • Adapted explanations dynamically to match student style
  • Used generated study plans to drive structured tutoring

RAG, Evaluation & Safety

  • Integrated RAG pipelines with vector DB retrieval
  • Applied Ragas for groundedness and faithfulness
  • Implemented strict acceptance policies and content moderation
  • Ensured safety for under-18 users

Outcome

  • Delivered a highly personalized and safe tutoring experience
  • Achieved major latency and cost improvements in production LLM workflows
PythonLangGraphLLMsRAGRagasPrompt EngineeringMultithreading

Enterprise Data Modeling & Data Mart Architecture

Jun 2023

Focus Area: Data Engineering (Foundational)

Objective

Design scalable data models and data marts to support analytics, reporting, and ML workloads.

Responsibilities & Contributions

  • Designed complete dimensional data models aligned with KPIs
  • Built data marts using enterprise modeling standards
  • Translated raw data into structured analytical schemas
  • Ensured consistency, scalability, and maintainability
  • Coordinated with analytics teams to standardize metrics

Outcome

  • Delivered a robust analytical foundation for reporting
  • Improved metric consistency and trust in outputs
Data ModelingDimensional ModelingData MartsSQLBusiness KPIs

Incremental Data Pipelines, ADF Orchestration & SQL Optimization

Nov 2023

Focus Area: Data Engineering & Analytics (Production Systems)

Objective

Build and optimize production-grade ETL pipelines with reliable incremental loads and predictable performance.

Responsibilities & Contributions

  • Designed ADF pipelines for full and incremental data loads
  • Ensured stable ingestion for large datasets
  • Built complex SQL stored procedures for reporting logic
  • Performed deep SQL performance engineering:
  • Replaced heavy CTE logic with optimized temp tables
  • Reduced redundant temp tables via join re-engineering
  • Improved execution plans for high-volume workloads
  • Enabled Power BI paginated reports with robust backend SQL

Measurable Impact

  • Reduced execution time from ~40 min → ~21 min for 1M-record workloads
  • Improved data freshness and lowered compute cost
  • Delivered stable SLAs for reporting pipelines
Azure Data FactorySQL OptimizationIncremental LoadsPerformance TuningPower BIDatabricksPySpark

Projects

Mental Health Relapse Prediction System

Healthcare / Mental Health

Estimate relapse risk and surface contextual clinical guidance

— Built an ML pipeline using PHQ-9, DSM-5, patient journals, and clinician notes
— Performed sentiment, frequency, and contextual feature extraction
— Modeled relapse risk as a continuous score
— Evaluated multiple approaches using cross-validated metrics
— Selected LSTM-based models for temporal pattern learning
— Added a RAG-based intervention layer aligned with medical best practices
— Deployed using Flask + React with Docker and real-time DB sync

Diabetes Risk Prediction & Monitoring System

Preventive Healthcare

Predict diabetes risk and provide contextual interventions

— Developed a questionnaire-driven chatbot to collect KPIs
— Triggered ML-based risk prediction models
— Implemented scheduled retraining with checkpointing and best-model selection
— Built a two-tier Flask architecture with containerized deployment
— Integrated RAG-based health interventions
— Deployed as an internal application

LLM-Based Resume Screening System

HR / Talent Screening

Automate resume shortlisting

— Built a resume screening system using vector databases and RAG
— Indexed resumes and job descriptions for semantic retrieval
— Designed similarity-based scoring for candidate ranking
— Reduced manual screening effort and improved consistency

Education

CMR Institute of Technology — Hyderabad, India

Bachelor of Technology (Mechanical Engineering)

2019 – 2023 8.46 / 10.0

Certifications & Recognition

Star of the Quarter

Cognine Technologies

March 2025

Databricks Generative AI Fundamentals

Databricks

August 2024

View Credential