Nagendra Manthena
Open to work2 years experienceAI Engineer
AI Engineer building agentic AI platform components and RAG workflows.
Standing Rank
Rank Not Available
Developer Badges
No badges earned yet
Skills & Technologies
49 skillsWork Experience
Full-Stack AI Engineer
Tata Consultancy Services (TCS)
- Engineered an Agentic Retrieval-Augmented Generation (RAG) pipeline using LangGraph and FastAPI to automate application lifecycle management and data quality resolutions for the GEAR team, replacing manual human-in-the-loop workflows.
- Extracted, parsed, and vectorized enterprise data from 5,000+ ServiceNow requests and 120+ Confluence pages using Docling and ChromaDB to build a centralized semantic knowledge base.
- Automated query resolutions for internal application service owners, reducing overall manual IT support ticket flow by 60%.
- Designed and deployed REST APIs in C#/.NET 8 on AWS EC2, serving enterprise-scale workloads; introduced Redis caching to reduce average response latency by ~40%.
- Diagnosed OLTP query bottlenecks by designing an ETL pipeline to incrementally sync ServiceNow data into a columnar data warehouse, cutting dashboard query time by ~60%.
- Integrated RabbitMQ for asynchronous inter-service messaging, decoupling 3+ microservices and improving fault tolerance.
Projects
Health Insurance AI Copilot
View Project- Built a stateful 10-node LangGraph agent with automated intent classification, query decomposition, and a self-critique feedback loop to resolve complex multi-hop health insurance queries end-to-end.
- Implemented PII detection/redaction and output safety-check guardrails within the agent pipeline, and used Pandas to clean and transform raw CSV source data prior to embedding and ingestion.
- Implemented a hybrid retrieval pipeline combining ChromaDB vector search, BM25 lexical search, a BGE Reranker, and a NetworkX Knowledge Graph (1,360+ nodes, 10,000+ edges) for high-precision document and provider lookups.
- Captured observability telemetry (query intent, retrieval latency, cache hit rates) enabling sub-second aggregation over millions of events; added Mem0 session memory and Redis semantic cache for sub-millisecond repeat responses.
- Containerized the full system with Docker – FastAPI backend and Streamlit frontend served via an Nginx sidecar, with end-to-end observability via LangSmith execution tracing.
FlowTalk
View Project- Designed a real-time voice assistant supporting browser and phone (Twilio) interfaces via a low-latency STT → LLM → TTS pipeline using Groq Whisper, Llama 3.3 70B, and Sarvam TTS.
- Built Twilio Media Streams integration with a custom G.711 µ-law codec in Go; leveraged Goroutines and Channels for a non-blocking concurrent pipeline with instant interruption handling – cancelling in-flight LLM/TTS requests when the user speaks.
- Optimized end-to-end latency via sentence-level TTS streaming and LLM tool calling (web search, weather, calculator, reminders) with live STT/TTFT/E2E telemetry in the UI.
LocalMind AI
View Project- Exported a PyTorch embedding model to ONNX and accelerated inference via ONNX Runtime + DirectML, enabling hardware-agnostic GPU acceleration across NVIDIA, AMD, Intel, and Qualcomm GPUs with zero cloud dependency.
- Engineered a stateful LangGraph agentic workflow with intent classification, FAISS neural retrieval, LLM answer generation, and a self-critique loop (relevance + groundedness + completeness scoring) that auto-re-retrieves when confidence < 0.6.
- Built a multi-stage Docker image (lean runtime ~1.2 GB) for clean Windows on-device deployment; exposed FastAPI endpoints returning structured JSON with per-answer confidence scores.
Leaderboard Standings
Leaderboard Position Pending
Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.
Assessment Highlights
Assessments Not Completed
Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.
AI Collaboration Score
AI Collaboration Score Pending
Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.
Role Compatibility Profile
Role Compatibility Analysis Pending
Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.
Achievements
Competitive Programming Achievements
LeetCode Knight – Global Top 5%, Max Rating 1909 | 3,000+ DSA problems solved | CodeChef 1606 | Codeforces 1447
About Details
Professional Bio
AI Engineer building agentic AI platform components, agents, tools, and observability-first RAG workflows. LeetCode Knight (top 5% globally) with expertise in Python, LangGraph, and production LLM/agent systems.
B.Tech in Electronics and Communication Engineering in Electronics and Communication Engineering
National Institute of Technology Andhra Pradesh (2020 - 2024)