Ishaan Sachdeva
Menu

Full-Stack Software Engineer and Technology Strategist

Ishaan
Sachdeva

Software engineer and technology strategist with 2+ years of experience across enterprise tech and consumer-facing platforms at Micron and Amazon. I build AI-powered applications end to end.

View projects Résumé PDF Get in touch
Portrait of Ishaan Sachdeva
Brooklyn, NY
Focus
Full-stack, AI products, strategy
Current
Technology Analyst, NYU Stern EFL
Base
Brooklyn, NY
Languages
English, Hindi, French

30%

lower device downtime, 150+ Amazon centers

75%

shorter time-to-productivity at Micron

sub-1s

firmware activation SLA for Dell and HP

73%

directional accuracy, forecasting system

01

About

I move between shipping code and shaping the business case behind it, from enterprise SSD firmware to AI market strategy.

Outside of work I was a Mitacs Globalink Research Fellow at ÉTS Montréal, and I write articles on AI and software engineering for Bootcamp on Medium and dev.to. I am pursuing an MS in Engineering Management at NYU, expected May 2027.

See education See experience

02

Projects

Each sheet shows what the project does. Source code is on GitHub.

Phone chat screen: a mood-based reply with menu categories, and labels for Gemini 2.5 Flash and Node.js with Express Menu · Chat · Cart add those Mood: comfort 2 dishes Mains Drinks Desserts Gemini 2.5 Flash Node.js Expo app

Conversational AI app

Intelligent Bistro

An AI-powered restaurant ordering app that replaces keyword search with natural-language ordering and mood detection. Gemini 2.5 Flash drives the chat assistant and personalized recommendations, with a Node.js and Express backend and a React Native (Expo) frontend.

  • Detects one of 8 moods from free text and opens with a tailored message plus 2–3 matching dishes.
  • React Native
  • Expo
  • Node.js
  • Express
  • TypeScript
  • Gemini 2.5 Flash
  • Zustand

Source intelligent-bistro

A water glass filled to a level, with labels for on-device storage, no ads and Google Play On-device storage No ads Google Play

Consumer mobile app

Hydrate Yourself

A consumer health utility built after a competitive analysis of wellness apps surfaced user pain points around advertising and privacy. It is a zero-commercialization Flutter app with fully on-device storage, published on Google Play.

  • Shipped to the Google Play Store with a 5.0 rating.
  • Flutter
  • Dart
  • On-device storage
  • Google Play

Source Google Play

Chart of news sentiment bars beneath a price line, with a label for the agentic evaluation loop S&P 500 price news sentiment agentic model evaluation

AI research

Financial Forecasting System

NYU graduate research project: a multimodal forecasting platform that pairs financial-news sentiment with S&P 500 price data. Agentic workflows handle automated model evaluation.

  • Achieved 73% directional accuracy.
  • RAG
  • Agentic workflows
  • Sentiment analysis
  • S&P 500 price data
Several MCP servers feed one gate with OAuth, RBAC, rate limit, schema and audit checks, and leave as one server notesinventoryother servers OAuthRBACrate limitschemaaudit log one MCP server

AI security tooling

mcpgate

A security gateway for the Model Context Protocol that fronts many MCP servers as one. Every tool call passes through OAuth bearer validation, per-client RBAC, rate limits, schema validation and a tamper-evident audit log.

  • 135 tests with 95.0% branch coverage.
  • Python
  • MCP
  • OAuth 2.1
  • JWT/JWKS
  • RBAC
  • Rate limiting

Source mcpgate

An LLM agent passes a policy gate to tools, with a human approval branch and a hash-chained audit log below LLM agent policy tools human approval default deny, taint tracking hashhashhashhash audit log, each entry chained

AI security tooling

agentguard

A capability-based permission engine for LLM agent tool calls. YAML policies, argument constraints, taint tracking and human approvals sit between the model and its tools, and every decision goes into a hash-chained audit log.

  • Ten-stage decision procedure with default-deny, least-privilege grants per task.
  • Python
  • YAML policies
  • Ollama
  • Taint tracking
  • Audit logging

Source agentguard

Paired bars for base and head Terraform plans, with a dashed CI gate threshold line Monthly cost diff base head CI gate

Developer tools

planprice

An offline Go CLI that prices Terraform plans and diffs monthly cost between two plans from a pinned AWS price snapshot. It outputs PR-ready Markdown and can gate CI on cost increases.

  • CI gate exits with code 3 when the monthly increase exceeds a set dollar or percent threshold.
  • Go
  • Terraform
  • AWS
  • CLI
  • CI/CD

Source planprice

A pipeline graph: load, split, featurize, train and evaluate, with train marked stale and evaluate pending loadsplitfeaturizetrainevaluate up-to-datecachedup-to-datestalepending content-addressed cache keys

Developer tools

dagcache

A content-addressed, parallel pipeline runner for reproducible ML workflows. It hashes each stage’s code, parameters and input data into a cache key, reruns only what changed, and explains why a stage is stale.

  • About 4,600 lines of typed Python with 165 tests.
  • Python
  • YAML
  • SHA-256 hashing
  • S3
  • Parallel execution

Source dagcache

A mood heatmap grid beside an audio-feature radar chart and a genre trend line mood heatmap genre evolution audio features

Data platform

Spotify Listening Analytics

A data engineering platform that pulls personal Spotify listening history, playlists and audio features via OAuth 2.0. It builds a dimensional model in Snowflake with SCD Type 2 tracking and serves a Streamlit dashboard.

  • 13 unit tests cover OAuth, pagination, sessionization and rate-limit handling.
  • Python
  • Snowflake
  • dbt
  • Streamlit
  • Airflow
  • Spotify API

Source spotify-listening-analytics

Pipeline flow from Open-Meteo through a Great Expectations gate to dbt and BigQuery, with a halt and alert branch hourly, Prefect Open-Meteo validate Great Expectations dbt BigQuery halt + alert

Data platform

Weather Data Pipeline

A weather data pipeline that collects hourly observations for 50+ global cities from the Open-Meteo API. It validates each batch with Great Expectations and transforms data with dbt into analytics-ready BigQuery tables.

  • Quality gates halt the pipeline and send a Slack alert when validation fails.
  • Python
  • BigQuery
  • dbt
  • Great Expectations
  • Prefect
  • Open-Meteo API

Source weather-data-pipeline-great-expectations

More sheets

  • Two learning curves, a random baseline dashed and an active strategy solid, with a target line

    labelbudget

    An active-learning benchmark that reports labels needed to reach a target accuracy with variance across many seeds, beside a prominent random-sampling baseline.

    Source labelbudget

  • A fatigue decay curve against a seasonal wave, with a dashed refresh marker

    fatigue

    Models ad creative fatigue as a hazard function of cumulative per-user frequency rather than calendar time, and predicts a refresh date per creative with a confidence band.

    Source fatigue

  • Two overlapping page bars inside a dashed outline, with scattered pairs faded out around them

    cannibal

    Search cannibalisation detection that tests page pairs with multiple-testing correction, so the output is a short list of real competing pairs instead of hundreds of noisy ones.

    Source cannibal

  • Three stacked data layers, from raw files to partitioned Parquet to analytics tables

    nyc-taxi-data-lake-gcp

    Processes the NYC TLC Trip Record Dataset (50M+ rows/month) with Dataproc + PySpark, writes optimized Parquet to GCS with Hive-style partitioning, serves via BigQuery external tables, and transforms with dbt.

    Source nyc-taxi-data-lake-gcp

03

Experience

  1. Sep 2026 – Present

    New York, NY

    Technology Analyst

    TEAMSEC | NYU Stern Endless Frontier Labs

    Securitization infrastructure venture at NYU Stern's Endless Frontier Labs, where I turn AI and data technology into market-entry strategy and product roadmaps.

    • Own the US market-entry workstream for an AI platform, building a bottom-up TAM/SAM/SOM model to size the opportunity and recommend a target segment.
    • Evaluate the platform's AI and data architecture against market requirements, assess data readiness, and translate gaps into a prioritized roadmap for the founder and product team.
    • AI
    • Data architecture
    • TAM/SAM/SOM market sizing

  2. Jul 2023 – Jul 2025

    Bangalore, India

    Senior Software Engineer

    Micron Technology

    Enterprise Solid State Drive (SSD) firmware engineering, delivering for Fortune 500 clients. Promoted from Software Engineer within 24 months.

    • Held a sub-1-second SLA window on firmware activation for enterprise SSDs (4TB to 128TB) shipped to Dell and HP, enabling zero-disruption upgrades in live production environments.
    • Cut the open defect backlog by 10% and improved release-date predictability by leading root-cause diagnostics with product validation teams and client engineers on integration failures.
    • Recovered ~2 weeks of recurring manual engineering effort by designing a centralized test automation framework standardized across 100+ concurrent setups.
    • Onboarded and mentored 10 engineers, building the onboarding curriculum and knowledge base from scratch and reducing time-to-productivity by 75%.
    • Synchronized offshore delivery with US product leadership across a 12-hour gap, translating executive roadmap priorities into local engineering scope.
    • SSD firmware
    • Test automation

    Earlier: Software Engineer, from Jul 2023

  3. Jan 2023 – Jun 2023

    Gurgaon, India

    Software Development Engineer Intern

    Amazon

    Fulfillment Technologies internship working on internal web applications and automation dashboards.

    • Drove a 30% reduction in device downtime across 150+ fulfillment centers by gathering requirements with senior operations leaders and redesigning automation dashboards around a new error-to-success ratio metric.
    • Shipped feature enhancements on internal web applications, owning the full cycle of requirement analysis, testing, risk evaluation and phased production rollout in an Agile cross-functional team.
    • AWS EC2
    • AWS S3
    • AWS Lambda
    • AWS SNS
    • AWS SQS
    • AWS CloudFormation

04

Skills

Web & Mobile

  • TypeScript
  • React
  • React Native
  • Expo
  • Node.js
  • Express
  • Flutter
  • Dart

Programming & Data

  • Python
  • SQL
  • Advanced Excel
  • Power BI

AI & Technical

  • Generative AI
  • Machine learning
  • LLM and agentic workflows
  • RAG
  • LangChain
  • Prompt engineering
  • Model evaluation
  • Process automation
  • Data privacy
  • AI-assisted development (Claude Code)

Cloud

  • AWS EC2
  • AWS S3
  • AWS Lambda
  • AWS SNS
  • AWS SQS
  • AWS CloudFormation

Strategy & Delivery

  • Structured problem-solving
  • Use case definition
  • Business case development
  • Roadmapping
  • Market sizing (TAM/SAM/SOM)
  • KPI and metric definition
  • Stakeholder management
  • Cross-functional delivery
  • Agile
  • Low-code prototyping

05

Education

Sep 2025 – Expected May 2027

In progress

New York University

MS, Engineering Management

  • GPA 3.80 / 4.00; GMAT Focus 655
  • Coursework: Financial Strategy, Economics, Machine Learning for Business
  • Course Assistant: Corporate Finance

Aug 2018 – Jul 2023

IIIT Bangalore

Integrated B.Tech + M.Tech, Electronics and Communication Engineering (ECE)

  • GPA 3.50 / 4.00
  • Top 1% nationally in JEE

06

Get in touch

Interested in full-stack engineering, AI products or technology strategy? Email or LinkedIn is the best way to reach me.