import json import re from collections import Counter # ── Raw resume text ─────────────────────────────────────────────────────────── RESUME_TEXT = """ NILESH HANOTIA AI Implementation Lead | POC-to-Production Specialist | Enterprise AI Deployment Ahmedabad, India | +91-9925537229 | nilesh.hanotia@outlook.com CERTIFICATIONS AWS Certified AI Practitioner | Amazon Web Services PROFESSIONAL SUMMARY AI implementation leader with 7+ years driving enterprise AI deployments from first discovery call through production go-live. Specialise in POC design and execution, stakeholder alignment, and translating complex LLM/AWS architectures into measurable business outcomes. At Armakuni, serve as the bridge between technical delivery teams and enterprise clients — owning the full 11-stage POC lifecycle: project allocation, internal/external kick-offs, weekly governance, execution, testing, delivery, and path-to-production planning. Independently delivered 6+ AI products generating $5M+ combined revenue, with proven impact in healthcare, legal, analytics, and automation domains. CORE COMPETENCIES AI Implementation & POC Delivery: End-to-end POC lifecycle ownership, POC→Production roadmapping, enterprise AI adoption strategy, deployment governance, MVP scoping, client validation Technical Fluency: LLM pipelines & AI workflows, AWS services (Lambda, Bedrock, API Gateway, DynamoDB, CloudWatch), serverless architecture, API integrations, solution architecture collaboration, logging & observability (Langfuse, Langsmith) Stakeholder & Delivery Leadership: Executive alignment, discovery workshops, SOW-to-delivery conversion, weekly PSR reporting, risk & escalation management, account expansion strategy, cross-functional team coordination PROFESSIONAL EXPERIENCE Customer Solution Consultant (CSC) — AI Implementation — Armakuni 2025 – Present Own end-to-end delivery of AI POCs and rapid engagements (5–8 weeks) across enterprise clients in healthcare, legal, and automation sectors. - Drive the full 11-stage POC lifecycle: project allocation, SOW review, internal/external kick-offs, planning, weekly governance, execution oversight, testing & client validation, delivery, and path-to-production. - Convert Statements of Work into executable delivery plans with milestones, sprint cadences, RACI matrices, and risk registers. - Serve as primary bridge between developers, solution architects, DevOps, and enterprise client stakeholders — ensuring technical decisions align with business outcomes. - Lead weekly executive reporting (PSR), risk management, and scope control; escalate proactively across a structured L1→L3 escalation matrix. - Develop deep working knowledge of AWS services (Lambda, Bedrock, API Gateway, CloudWatch) and LLM pipeline architecture to engage credibly with both technical teams and client leadership. - Create and maintain technical documentation — process flows, README files, test cases — to ensure production readiness from day one of execution. - Facilitate path-to-production workshops with clients, mapping POC learnings to scalable architecture, compliance requirements (GDPR, HIPAA), and business value expansion plans. - Identify account expansion, cross-sell, and upsell opportunities, contributing to long-term Armakuni account growth strategy. AI Product & Implementation Lead — Independent Consultant 2021 – 2025 Delivered AI products end-to-end — from problem discovery and architecture design through deployment, user adoption, and ongoing iteration — serving 10,000+ users across multiple industries. - Designed and deployed 6+ production AI systems generating $5M+ combined revenue across healthcare, legal, enterprise analytics, and automation domains. - Led full AI product lifecycle: stakeholder discovery, requirements definition, LLM workflow design, MVP build, UAT, and phased production rollout. - Reduced manual client workflows by 80-90% through targeted AI automation, directly measurable in client operations. - Translated complex enterprise business needs into implementable technical architectures — working hands-on across AWS, API integrations, and AI/LLM pipelines. - Built and validated AI products with real users, iterating rapidly based on adoption data and feedback to drive retention and expansion. - Simultaneously built independent projects including real-time trading systems on AWS (Lambda, DynamoDB, API Gateway, Bedrock, CloudWatch) — deepening hands-on serverless and AI infrastructure expertise. EARLIER CAREER Product Analyst — Cignex Datamatics 2020 – 2021 Business Product Analyst — Cygnet Infotech 2019 – 2020 Associate Product Manager — Aimdek Technologies 2017 – 2019 TECHNICAL TOOLING Cloud & AI: AWS (Lambda, Bedrock, API Gateway, DynamoDB, CloudWatch, CloudFront, CDK) | LLM Pipelines | Langfuse | Langsmith Dev & Data: Python | SQL | APIs | Docker | Tableau | Mixpanel Delivery & Design: Jira | Zoho Projects | Figma | Panda.doc | Agile/Scrum EDUCATION Post Graduate Diploma in Business Administration — St. Kabir Institute Bachelor of Commerce — IGNOU """ # ── Step 1: Build character-level vocabulary ────────────────────────────────── def build_vocab(text): # Count frequency of every character char_counts = Counter(text) # Sort by frequency (most common first) sorted_chars = sorted(char_counts.items(), key=lambda x: -x[1]) # Assign ID to each unique character stoi = {ch: idx for idx, (ch, _) in enumerate(sorted_chars)} itos = {idx: ch for ch, idx in stoi.items()} return stoi, itos, char_counts, sorted_chars # ── Step 2: Encode full text ─────────────────────────────────────────────────── def encode(text, stoi): return [stoi[ch] for ch in text if ch in stoi] # ── Step 3: Extract key terms from resume ───────────────────────────────────── def extract_keywords(text): text_lower = text.lower() skills = ["python", "aws", "sql", "docker", "llm", "bedrock", "lambda", "api", "langfuse", "langsmith", "tableau", "figma", "jira", "cdk"] education = ["bachelor", "post graduate", "diploma", "commerce", "business administration"] experience = re.findall(r'(\d+)\+?\s*years?', text_lower) locations = ["ahmedabad", "india"] roles = ["ai implementation", "product manager", "consultant", "analyst", "product lead", "solution consultant"] found_skills = [s for s in skills if s in text_lower] found_education = [e for e in education if e in text_lower] found_roles = [r for r in roles if r in text_lower] return { "skills" : found_skills, "education" : found_education, "experience": experience, "location" : locations, "roles" : found_roles } # ── Main ─────────────────────────────────────────────────────────────────────── if __name__ == "__main__": print("=" * 55) print(" VOCABULARY BUILDER — NILESH HANOTIA RESUME") print("=" * 55) stoi, itos, char_counts, sorted_chars = build_vocab(RESUME_TEXT) print(f"\nTotal characters (with repeats) : {len(RESUME_TEXT)}") print(f"Unique characters : {len(stoi)}") print(f"Vocab size : {len(stoi)}") print("\n── Top 15 most frequent characters ──") for ch, count in sorted_chars[:15]: display = repr(ch) print(f" {display:6s} → ID {stoi[ch]:3d} | appears {count:4d} times") print("\n── Full Character → ID mapping ──") for ch, idx in stoi.items(): display = repr(ch) print(f" {display:6s} → {idx}") # Encode a sample sample = "AI Implementation Lead with AWS and Python skills" encoded = encode(sample, stoi) print(f"\n── Sample Encoding ──") print(f" Text : {sample}") print(f" Encoded : {encoded}") # Extract keywords print(f"\n── Extracted Keywords from Resume ──") keywords = extract_keywords(RESUME_TEXT) for key, values in keywords.items(): print(f" {key:12s}: {values}") # Save vocab to JSON vocab_data = { "stoi" : stoi, "itos" : {str(k): v for k, v in itos.items()}, "vocab_size": len(stoi), "keywords" : extract_keywords(RESUME_TEXT) } with open("vocab.json", "w") as f: json.dump(vocab_data, f, indent=2) print(f"\n✅ vocab.json saved — {len(stoi)} characters mapped") print("✅ Ready for Step 2: Tokenizer")