// Boston, MA · Massachusetts, USA
> initialize(neural_engine, llm_stack, data_ops)
✓ LLM runtime loaded → GPT-4 / Claude / Mistral
> connect(RAG_pipeline, Pinecone, PostgreSQL)
✓ Vector store synced · 99.9% uptime confirmed
> loading(portfolio_v3, theme="neural_garden")
✓ Portfolio ready · Let's connect
Mounting experience
AI / ML Engineer · Boston, MA
Aditi
Mishra

Building production AI systems that think, reason, and act — from fine-tuned LLMs to enterprise RAG pipelines at scale.

LangChain RAG Pipelines Fine-Tuning · LoRA PySpark Pinecone RAGAS · TruLens Kubernetes MCP Protocol
View Projects ↓ Resume Let's Connect
aditi_mishra · profile.json
$ →query_profile --full

name: "Aditi Mishra"

role: "AI / ML Engineer"

current: ScaleUp Labs · Boston

yoe: "3+ years production AI"

location: "Boston, MA"

$ →get_impact --verified

inference_cost: ↓ 70% at ScaleUp Labs

data_accuracy: 87% → 98.1% at IBM

students_mentored: 1,000+ at Google DSC

pipeline_uptime: 99.9% over 18 months

$ →_
70% Inference cost reduced at ScaleUp Labs· 90% Manual data entry eliminated via OCR+LLM pipeline· 99.9% Pipeline uptime over 18 months at IBM· 1,000+ Students mentored through AI/ML at Google DSC· 35% Retrieval accuracy improvement via optimized RAG· 87→98% Data accuracy improvement at IBM· 70% Inference cost reduced at ScaleUp Labs· 90% Manual data entry eliminated via OCR+LLM pipeline· 99.9% Pipeline uptime over 18 months at IBM· 1,000+ Students mentored through AI/ML at Google DSC· 35% Retrieval accuracy improvement via optimized RAG· 87→98% Data accuracy improvement at IBM·
01 credentials / education

Academic
foundation

🎓 Most Recent · 2025
2024 – 2025 · Graduate
Master of Science in Data Science
Clark University · Worcester, Massachusetts
3.7
GPA · Dean's List
2018 – 2022 · Undergraduate
B.Tech in Electronics Engineering
Dr. Abdul Kalam Technical University · India
3.4
GPA · Honors
→ Industry Certifications
☁️
AWS Certified Cloud Practitioner
⚡
Databricks Certified · Apache Spark 3.0
🐍
Machine Learning with Python · IBM
🪟
Microsoft MTA · Programming Using Python
02 work_history / experience

Where I've
built AI

🚀
ScaleUp Labs · Boston, MA
AI-first startup building enterprise automation tooling
May 2025 – Present● Current
AI Engineer
⚡ Cut inference costs by 70% — rebuilt RAG system (LangChain + Pinecone) with re-ranking, query compression, and chunking optimization, reducing monthly LLM spend from $14K → $4.2K while improving retrieval accuracy by 35%.
📄 Automated 90% of manual document processing — built OCR + LLM pipeline (Tesseract + GPT-4) that converts unstructured PDFs into structured PostgreSQL records, eliminating a 3-person manual review team.
🔗 Shipped enterprise MCP interface — designed Model Context Protocol layer enabling secure, auditable LLM ↔ enterprise data connectivity with role-based access controls.
📊 Reduced hallucination rate by 42% — built LLM evaluation harness using RAGAS and custom metrics; integrated semantic personalization layer for response grounding.
→ stack
LangChainPineconeRAGMCPTesseract OCRPostgreSQLLoRA/PEFTRAGASHugging Face
🔷
IBM · India
Global technology and consulting, data engineering division
Jun 2022 – Aug 20242 yrs 2 mos
Data Engineer
📈 Improved data accuracy from 87% to 98.1% — built automated validation pipeline using Great Expectations with 200+ rules, enabling the ML team to ship models 3× faster by eliminating manual data cleaning sprints.
🏗️ Maintained 99.9% uptime across 40+ daily ML pipeline jobs — re-architected PySpark ETL on Hadoop with Docker + Kubernetes orchestration and Jenkins CI/CD, cutting incident response from 4 hours → 22 minutes.
🚀 Reduced feature engineering time by 60% — redesigned data ingestion to publish real-time feature streams via Apache Kafka, replacing nightly batch jobs that blocked 6 downstream ML experiments daily.
→ stack
PySparkHadoopApache KafkaGreat ExpectationsDockerKubernetesJenkinsPostgreSQL
🎓
Google Developer Student Club · India
University-level AI/ML community leadership
Sep 2021 – Jun 2022Lead
Lead — AI / ML / Data Science
👩‍🏫 Mentored 1,000+ students through hands-on NLP & Computer Vision projects, achieving an 85% end-to-end project completion rate on GCP deployments — highest in the club's history.
📚 Designed and delivered 12-session ML curriculum covering model training to cloud deployment; curriculum adopted by 2 other chapters in the following semester.
→ stack
TensorFlowPyTorchGoogle CloudNLPComputer Vision
verified impact / by the numbers

Real numbers,
real systems

LLM Inference Cost
0%↓
$14K → $4.2K/month via RAG optimization at ScaleUp Labs
Source: ScaleUp Labs · 2025
Data Accuracy
0%
87% → 98.1% via automated Great Expectations validation at IBM
Source: IBM · 2022–2024
Pipeline Uptime
0%
Across 40+ daily ML jobs over 18 months at IBM via K8s + CI/CD
Source: IBM · 2022–2024
Students Mentored
0+
85% project completion rate · Curriculum adopted by 2 chapters
Source: Google DSC · 2021–2022
03 technical_skills / expertise

What I
build with

3+ years applying these tools in production — shipping RAG systems, ETL pipelines, and LLM evaluations at IBM and ScaleUp Labs.

🧠
LLM & GenAI
Production · 2 years
RAG Pipelines
LangChain / LlamaIndex
Fine-Tuning · LoRA / PEFT
Prompt Engineering
LLM Eval · RAGAS · TruLens
Hugging Face Transformers
Used in production at
ScaleUp LabsPersonal Projects
⚡
Data Engineering
Production · 3 years
Python · SQL
PySpark · Apache Spark
ETL Pipeline Design
PostgreSQL · Data Modeling
Apache Kafka
Data Quality · Great Expectations
Used in production at
IBM · 2 yrsScaleUp Labs
☁️
Cloud & MLOps
Production · 2.5 years
AWS · GCP · Azure
Pinecone · FAISS · Weaviate
Docker · Kubernetes
MLOps · Model Serving
Jenkins · CI/CD
Terraform · IaC
Used in production at
IBM · K8s clusterScaleUp Labs
04 projects / deployed_systems

AI I've
built & shipped

Artha Savvy · v1.2
⌥ GitHub ↗ Demo
Artha Savvy
AI Finance Agent
Conversational agent that analyzes personal finances and generates risk-aware plans via LLM reasoning over private financial documents — zero data leaves the user's environment.
→ Architecture
User query → Mistral-7B (self-hosted) → FAISS retriever over PDF corpus → Re-ranker → Plan generator → Structured JSON → React UI. All inference on-device.
→ Stack
LangChainMistral-7BFAISSReactFastAPI
80%Planning tasks automated
100%On-device · zero data egress
1.2sAvg response latency
artha-savvy.vercel.app · Finance Chat
▶
Watch Demo · 2m 14s
EU AI Act Scanner · v1.0
⌥ GitHub ↗ Demo
EU AI Act
Compliance Scanner
Automated auditing system that classifies AI services against EU AI Act risk tiers, maps requirements, and generates structured compliance reports — replacing a 2-week manual audit process.
→ Architecture
Service descriptor → GPT-4 risk classifier → EU Act requirement mapper → Gap analysis → PDF report generator. Processes 1 service in <4 minutes vs. 2-week manual baseline.
→ Stack
OpenAI GPT-4LangChainReportLabFastAPIStreamlit
60%Audit time reduced
<4 minPer service vs. 2 weeks
3 tiersRisk classification
eu-scanner.streamlit.app · Compliance Report
▶
Watch Demo · 1m 48s
RAG Eval Harness · Open Source
LLM Evaluation · Developer Tool
⌥ GitHub ↗ Docs
RAG Evaluation Harness
Open-source toolkit for measuring RAG system quality across faithfulness, answer relevancy, context precision, and hallucination rate — built from production experience optimizing the ScaleUp Labs pipeline.
→ Architecture
Test dataset loader → Multi-metric evaluator (RAGAS, TruLens, custom embedding checks) → Regression detector → HTML/JSON report exporter. Plugs into any LangChain or LlamaIndex pipeline with 3 lines of code.
→ Stack
PythonRAGASTruLensDeepEvalLangChainPytestJinja2
42%Hallucination reduction at ScaleUp
6Evaluation metrics tracked
3 linesTo integrate into any RAG pipeline
system_design / architecture

How I architect
AI systems

Interactive diagrams of the production systems I've built. Each node is a real component — hover to inspect. Watch data flow in real time.

LLM / Model Layer
Data / Storage
Processing
Interface / Output
Evaluation
initiate_connection / contact

Let's build
something
intelligent.

Open to AI/ML Engineering roles and production AI collaborations. Let's build something remarkable together.

✉ aditi.mishra28@gmail.com in LinkedIn ⌥ GitHub ↓ Resume PDF